Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Spacematch: Using environmental preferences to match occupants to suitable activity-based workspaces

Spacematch: Using environmental preferences to match occupants to suitable activity-based workspaces The activity-based workspace (ABW) paradigm is becoming more popular in commercial oce spaces. In this strategy, occupants are given a choice of spaces to do their work and personal activities on a day-to-day basis. This paper shows the implementation and testing of the Spacematch platform that was designed to improve the allocation and management of ABW. An experiment was implemented to test the ability to characterize the preferences of occupants to match them with suitable environmentally- comfortable and spatially-ecient flexible workspaces. This approach connects occupants with a catalog of available work desks using a web-based mobile application and enables them to provide real-time environmental feedback. In this work, we tested the ability for this feedback data to be merged with indoor environmental values from Internet-of-Things (IoT) sensors to optimize space and energy use by grouping occupants with similar preferences. This paper outlines a case study implementation of this platform on two oce buildings. This deployment collected 1,182 responses from 25 field-based research participants over a 30- day study. From this initial data set, the results show that the ABW occupants can be segmented into specific types of users based on their accumulated preference data, and matching preferences can be derived to build a recommendation platform. Keywords: IoT - Internet of Things, Thermal comfort, Space utilisation, Flexible work arrangement, Activity-based workspaces 1. Introduction years have led them to rethink spatial density and utilization. In the past few years, rising corporate real estate (CRE) costs 1.1. The emergence of workplace flexibility and rapid changes in technology and nature of work have ren- dered inecient traditional modes of working where occupants In response to these challenges, new ways of working are are permanently designated a single work desk. Today, 37% evolving rapidly. These approaches aspire to simultaneously of all oce spaces are empty on any given workday [1], which balance operator’s cost and space saving demands with flexibil- equates to approximately 150 billion dollars annually in unused ity and comfort needs of employees through enabling occupant space globally [2]. These challenges are pushing building oper- mobility. Through most of these approaches, the occupant can ators to rethink occupant density and spatial utilization in work- work flexibly by choosing di erent spots within the workplace places. In a recent survey, while 95% of CRE professionals rather than being assigned a fixed desk as the one primary place believed that workplaces influence occupant productivity and of work. Once occupants are dynamic in the ways they use comfort, only one third measured that impact. Most others only space, it is easier to recapture underutilized spaces by operators considered traditional cost-based measures and metrics to quan- [2]. Workplace strategies of this kind are often referred to as tify workplace density and utilization [3]. Based on past stud- activity-based workspaces (ABW) or by other terms such as hot- ies, heavy reliance on cost-based measures generally results in desking, co-working, desk-sharing, flexible working, and oce operators taking away amenities that make occupants comfort- hoteling. Though each strategy varies slightly from the other, able and productive in cost-saving work environments. An ex- most promise benefits of improved spatial utilization and cost ample is the replacement of cubicles with benches to accommo- savings for operators while increasing overall comfort, choice, date a more significant headcount, removal of informal collabo- and control for occupants [4]. Equally, an extensive recent sur- ration spaces for more desks, or taking away of employee stor- vey of spaces utilizing one such strategy showed that the pri- age space all together [2]. Workplaces today face challenges mary motivations for occupants to work in such a workplace is as operators look to o set high rental costs by growing occu- because it allowed access to an inspiring work environment [5]. pant headcount within their existing footprint. Equally, rapid Understandably, an increase in the adoption of these concepts changes in technology and the nature of work in the past few can be seen in the growing co-working industry. Preprint submitted to Frontiers in Built Environment June 18, 2020 arXiv:2006.09570v1 [cs.CY] 17 Jun 2020 1.1.1. COVID-19 and its e ect on the future of work In response to these challenges, research in indoor occupant comfort has accelerated over the last twenty years. One of In early 2020, oce working habits of much of the world the many paradigm shifts has been the movement away from changed due to a global pandemic of a novel pathogen known traditional, physically-based deterministic models due to their as COVID-19. As there is no treatment or vaccine for this reported low accuracy (only 34%) across dozens of comfort virus, numerous community-driven mitigation strategies have studies in the past decades [18, 19]. Recent research e orts been deployed across the world. One of the most common is have progressed towards adaptive comfort models [20, 21, 22], the requirement for those that can work from home to do so which rely on human behavior. According to these models, dis- [6]. This exodus from oce spaces to the home has shown that comforting changes in the thermal environment are followed by such decentralization of oce work is possible, and desirable a behavioral change in people to restore comfort. Such actions in some situations. Going forward this forced push towards could include reducing individual activity levels or even open- home working will reinforce the need for corporate entities to ing a window. The main e ect of such models is to increase adopt more agile real-estate portfolios, with more flexibility and the range of conditions that designers can consider comfortable, a more distributed footprint to ease employees back into the for instance, naturally ventilated buildings in the tropics where workplace and cut commuting time to a single, large headquar- occupants have a higher degree of control over their thermal ters [7]. This digitization and rethinking of how people can environment. work could be a strong catalyst for the adoption of ABW-style Despite these advancements, even adaptive comfort models oce arrangements. follow a one-size-fits-all approach that ignores the personal as- 1.1.2. The limitations of ABW strategies pect of comfort; this is like expecting everyone to have sim- Despite the momentum towards ABW, there are a significant ilar preferences for food, music, or style; all subjective at- number of challenges that this strategy poses through its shift tributes of a person’s personality. Further work on adaptive in oce culture. A study focused on a sociological analysis of models is needed to identify comfort preferences on an individ- one approach showed a ”loss of everyday workspace ownership ual basis and not only based on the thermal conditions. Recent giving rise to practical and social tensions within the organiza- studies have shown that occupants exposed to the same condi- tion [8]”. Another study found lower than expected satisfaction tions could exhibit variations in environmental perception due with activity-based working environments due to rare switching to individual di erences in comfort preferences and personal- of di erent activity settings [9]. A recent study even found evi- ity [23, 24]. Researchers have addressed this through the de- dence of dehumanization as a result of ABW [10]. As organiza- velopment of personal comfort models that predict individual tions evolve, they are wary of ill-conceived applications which thermal comfort responses rather than the average response of a larger population [25] through leveraging machine learning may disrupt business and culture purely for cost savings of new techniques and the Internet of Things (IoT) technologies. This workplace strategies. These studies illustrate that there is much approach demonstrates a very high prediction accuracy, well improvement possible in the deployment of ABW strategies to beyond that of the traditional models [26]. help mitigate these downsides. 1.2. Connection to indoor environmental comfort in buildings 1.2.1. Building systems’ response to indoor environmental challenges In addition to space use allocation, indoor environmental comfort is at the forefront of building performance analysis. In addition to the work in comfort models, a large amount Occupant dissatisfaction with indoor environments has far- of research has also been done to improve the systems that re- reaching economic implications for workplaces. As people typ- spond to comfort needs. Various mechanical systems technolo- ically spend over 90% of their time indoors, indoor environment gies such as radiant systems and decentralized and personal- quality influences their comfort, performance, health, and well- ized ventilation attempt to address comfort problems. These being. Occupant dissatisfaction with indoor environments can contemporary innovations focus on the ability of a building to result in health impacts, absenteeism, and reduced productiv- adapt to its occupants by tracking them and modifying each per- ity [11]. Not only do enterprises today associate a majority son’s immediate personal climate to meet their individual needs of their costs (80-90%) to workers compensation and benefits [27]. [12, 13, 14] but as people typically spend more (about 90%) Personalized control system approaches have limitations as time indoors, the quality of the indoor environments influences the spatial resolution of most existing climate, lighting, and their comfort, performance and well-being at work. Continual noise control technologies does not have enough flexibility and occupant dissatisfaction with indoor environments can result responsiveness. Even more innovative, decentralized and per- in health impacts, absenteeism, and reduced productivity [11]. sonalized comfort systems are unable to create the response and A comparison of recent field studies from 467 air-conditioned resolution needed to practically create individualized comfort buildings containing 24,000 occupants showed between 30% zones for all occupants. Additionally, smaller and more decen- and 200% more cases of sick building syndrome symptoms than tralized systems create more maintenance tasks and complexity in the occupants of naturally ventilated buildings [15, 16]. An- within the building systems [28]. To meet these challenges of other survey in 2012, of 52,980 occupants in 351 oce build- decentralization, a balance between personalizing spaces and ings, found that 50% of the occupants were dissatisfied with maintaining the economies of scale that centralized systems their indoor environments [17]. provide could be achieved. Creating small zones of personal 2 comfort for all occupants should be used sparingly in particular there are three occupants with indoor comfort preferences: Per- space use types, while most other spaces can be conditioned to son A enjoys a dim, quiet, and moderately conditioned space, meet the needs of a subgroup of people. Person B is into a warmer and more active environment with a bit of noise and brightness, and Person C likes a cooler envi- ronment that is not dead silent. Perhaps these preferences are 1.2.2. Collecting human comfort feedback in buildings relatively consistent, or they could be dependent on the occu- Currently, researchers and building owners install a wide pant’s current activity or frame-of-mind. If these occupants’ range of IoT devices that measure various environmental condi- preferences are collected over time, the probability of nudg- tions such as light, noise, and particulates levels, in addition to ing these users to spaces in oces that match these specific the conventional temperature and humidity metrics. Although needs is higher. Also, there is the possibility of grouping peo- IoT sensors have become low cost and ubiquitous, the data from ple with similar preferences, which could be combined with these devices are often not utilized to their full potential. Com- systems control to create di erent zones with di erent types of fort models, even adaptive ones, only set thresholds to which comfort. Perhaps Person A, B, or C could be grouped with peo- sensor data points can be compared. The critical element in ple who have comfort personality type A, B, or C to improve putting these data in context is subjective and physiological both satisfaction and systems control. The goal of this study is feedback from people who inhabit the environment. Collecting to try to capture these tendencies using a web-based tool over a this type of data would empower more specialized and nuanced more extended period than a typical indoor comfort study. comfort models to be developed in a scalable way [29]. How- ever, the quantitative identification of individual di erences in comfort preferences and personality of people remains a sig- 1.3.1. Novelty of proposed approach nificant challenge in the field for researchers and practitioners [30]. There have been several recent e orts with the focus of using One way to tackle this situation is to use contemporary meth- technology to improve the ABW paradigm. One primary direc- ods of personal feedback collection such as structured surveys tion has been on the use of occupancy detection and prediction or interviews either online or oine, in-person or remote, which to characterize the use of ABW [33]. Additional work focuses would help increase the frequency and volume of building oc- on the use of human-sensor interaction to promote better deci- cupant feedback. However, such conventional methods have sions by occupants [34]. A previous study has used occupancy several shortcomings [31]. One major drawback of these meth- data to optimize the allocation of hot desk spaces using sim- ods is the lack of scalability. It is dicult to collect large sample ulation and occupancy sensors [35]. In the literature, there is data sets due to the administrative, financial, and other oper- a single case of designing a seat recommendation system that ational overheads associated with these approaches. Further- has initial e orts towards matching people to spaces that would more, other factors such as lack of knowledge (respondents best match their needs [36]. Therefore, the presented platform do not know the answer to a question, but answer it nonethe- and methodology is among the first examples of data collec- less), lack of motivation (respondents may not process ques- tion, characterization, and e orts towards a space recommen- tions fully) and failures in communication (survey questions dation system. The presented scope of work combines the use may be unclear or misunderstood) result in an increased risk of of clustering occupants based on their preferences gathered in biases and respondent heuristics in traditional survey responses a longitudinally-intensive method and characterizing a comfort [32]. As sensor adaptation in built environments continues to matching preference probability. These are techniques which grow, new technologies and modern data capabilities allow re- have not been found in the literature. searchers and practitioners to e ectively capture dynamic hu- man feedback e ectively. However, this approach also presents significant challenges in the collection, analysis, processing, 1.3.2. Organization of the paper and visualization of large data sets from building occupants. This study addresses each of the previously mentioned chal- There is a need for easy to use, scalable solutions that help iden- lenges: collecting larger volumes of personalized data useful tify and quantify occupant comfort preferences for operators as to occupants and operators, reducing the need for complex they move to ABW. and problematic personalized comfort systems, and impacting ABW by improving spatial utilization and occupant comfort. 1.3. Towards improving ABW by recommending the best loca- This recommendation system’s hypothesis is to test whether tion for an occupant based on their preferences certain groups of occupants can be segmented according to their This paper outlines a platform that improves indoor environ- comfort preferences and whether this segmentation is realistic mental satisfaction and ABW by allocating occupants to spaces in the context of an actual building case study. This paper shares that are the best match for their needs. We seek to test whether a the development and testing of the platform in the context of a longitudinally-intensive collection of indoor comfort data from case study implementation. In Section 2, we first illustrate an individual occupants can be used to assign each person to a cer- implementation with 25 research participants over a month in tain preference tendency type. The goal is to use these prefer- six flexible workspaces. Section 3 showcases the results from ence types to match that person to a space that could best meet this implementation and Section 4 discusses the interpretation their needs. For example, consider a simple situation in which of those results in the context of future work. 3 Figure 1: Overview of the platform user flow - the goal is to give flexible workspace occupants the ability to find spaces that meet their needs (1 and 2), give feedback about their comfort to develop a comfort personality type (3), and eventually provide train a model (4) comfort suggestions for subsequent uses (5) 2. Methodology implementation, as shown in Figure 3. The zones were split be- tween the two institutional buildings across three di erent lev- To test the user segmentation and space allocation hypothe- els. Each zone was strategically selected based on di erences sis, a progressive web application platform was developed for in location, floor level, number of desks, and zone accessibility. implementation in the SDE4 and SDE2 building on the cam- Further, di erences in window-to-wall ratios, zone orientation, pus of the National University of Singapore. The first step was and proximity to the nearby vehicular road and public areas of the development of the user flow of the platform. Figure 1 il- the institution lead to varying light and noise levels between lustrates the user in the case study who can find available open zones. workspaces, reserve them for use, and give comfort feedback to As shown in Figure 3, desks in each zone were arranged help train a model to predict which location would be best for in the proximity of fixed indoor environmental quality sensors them based on their past comfort feedback. To use the platform, measuring seven attributes in real-time: temperature, humid- the occupants can choose to use the work desk right away or re- ity, noise, light, carbon dioxide, volatile organic compounds, serve for later use. During use, the mobile application enables and presence. Each zone o ered flexible work desks in which occupants to quickly provide environmental comfort feedback each arrangement di ered slightly. In some zones, desks were for temperature, noise, and light variables, as shown in Figure arranged to promote collaboration; they were aligned so par- 2. For flexible workspace operators, it facilitates merging per- ticipants could face each other while working. In others, the sonalized environmental comfort data with other data streams arrangement enabled solitary, personal work. Each desk was such as indoor environmental quality, occupancy, and energy identified through a unique label containing the desk number, use data among others to optimize space, energy usage, and room name, and a QR code for connecting the desk to the occupant comfort by grouping users with similar preferences. progressive web application, as shown in Figure 4. During the study, participants were not assigned a specific desk and 2.1. Implementation were encouraged to alternate their desk usage between di er- ent zones rather than stick to a single zone. The experimental A case study was designed to test the platform in field condi- instruction were designed help to provide generality and eco- tions with 25 research participants over one month in April and logical validity to the collected data. May 2019. The participants were recruited from the current mix of students and sta , which were representative of regular users Each participant in the experiment used the interactive mo- of two selected institutional buildings. An ethics review was bile application to reserve and use work desks, as shown in Fig- submitted approved for the methodology of the study. A total ure 2. Participants could search for available work desks within of 36 desks in six di erent zones were identified for case study the two institutional buildings. Once they chose the building, 4 Figure 2: Overview of the application: (a) Find flexible working zones around campus, (b) Dashboard and information screen for more details regarding each zone, (c) Choose between options for use desk now or reserve desk for later, (d) Provide feedback for temperature, light and noise variables participants could progress to select the room and desk to use. At the start of the pilot study, a common onboarding session They were provided with more information about the room and for all participants was organized. During this session, goals, real-time indoor environmental quality through the Info and objectives, and the methodology for the study were discussed, Dashboard features of the application, as shown in Figure 2b. and the participants were onboarded to the interactive mobile The application provided options between using the desk right application. Using the platform, the research team demon- now or reserving it for later. To start using a desk, participants strated to the participants on how to find and book a desk and had to scan the desk QR code label using an in-built QR code to provide feedback during desk usage. Details such as partic- scanner in the application. A minimum of a two-hour time slot ipation schedule, timings, zone locations, and physical acces- for a work session was provided for each desk booking. Partic- sibility guidelines were also shared during this session. After ipants could also choose to extend their work sessions in two- this session, flexibility was provided for participants to use any hour multiples as per their requirements using the application. of the six zones between 8 am to 6 pm daily for the month dur- During use, the application prompted participants to provide ing the pilot study. Participants could also choose to participate environmental feedback for temperature, light, and noise levels in groups or individually based on their routine personal and through a three-point scale, as shown in Figure 2d. Prompts work preferences to ensure that there is no disturbance to the were configured such that the application nudged users to pro- typical flexible workspace environment. However, participants vide feedback at the start, finish, and once every half an hour of were encouraged to alternate their desk usage between di er- a typical two-hour work session. This miniature survey is a type ent zones and times of the day to provide variety, generality, of ecological momentary assessment, a method for longitudinal and ecological validity to the experimental findings. The test data collection pioneered in medicine and psychology [37] and participants gave individual feedback in a range of 40-100 total recently adapted more for environmental perception [38]. feedback points per person. 5 Figure 3: The six flexible working zones for case study implementation. The orange square in each zone layout represents the location of the indoor environmental quality sensors: (a) Layout and photo from Zone 1, (b) Zone 2, (c) Zone 3, (d) Zone 4, (e) Zone 5, (f) Zone 6, (g) Selection of the six zones across three floors between two separate institutional buildings The data from the users and fixed sensors were aggregated platform for data acquisition, storage, and error detection, as using a cloud-based, time-series database, which served as a shown in Figure 4. The combination of location-based user 6 Figure 4: Data exchange framework: (a) Desk QR code label, (b) Web-based mobile application, (c) IoT based indoor environmental quality sensors, (d) Time series data base comfort feedback and fixed environmental sensor data allowed grouping similar preference behavior. This e ort captures each clustering analysis for personalized comfort profiles of users. user’s behavior in their interaction with the system instead of These two data sources were merged through matching feed- the demographic, physiological, or environmental conditions back location (spatially localized through desk QR code label), variables that are typically addressed in environmental prefer- time of feedback collection (timestamp), and user ID (through ence studies. an in-built anonymous authentication method). 3.1. Discovering occupant personal comfort preference types 3. Results To cluster user preferences, unsupervised learning tech- niques were used to group the participants into cohorts with This section analyzes the environmental quality comfort similar feedback for temperature, light, and noise variables, as preferences for temperature, light, noise values from 25 re- shown in Figure 5. The analysis leads to identifying four dis- search participants in the pilot study. An unsupervised clus- tinct clusters based on di erences in preferences for tempera- tering technique is applied to segment participant comfort data, ture, light, and noise levels across participants. As shown in totaling 1,182 feedback points, into clusters based on similar Figure 5a, many participants are generally comfortable across behavior. This study focuses on a participant’s behavior based zones. However, some participants preferred cooler or warmer on their interaction with the system rather than on conventional environments. For participants in Hard to Predict cluster, variables in similar environmental preference studies, such as more extensive and more diverse data streams are needed to demographics, physiological, or environmental conditions. The understand their preferences better in the future. For visual emphasis is to apply an unsupervised clustering technique to comfort or light values related preferences, the clustering is the occupant data to segment the users who provide more than evenly spread between prefer dimmer and generally comfort- five feedback points into cohorts of similar behavior. This type able choices, as shown in Figure 5b. As can be observed, most of analysis focuses on the characterization of comfort prefer- participants would prefer a change in their light settings across ences in ways specific to each occupant, but generalizable by zones. As shown in Figure 5c, most participants were aurally 7 Figure 5: User comfort preference clustering based on: (a) thermal comfort feedback (degrees Celcius), (b) noise comfort feedback (dB), (c) light comfort feedback (lux) 8 comfortable with a few preferring a change in noise levels. Un- and easy-to-use measures can be employed without the loss of derstanding a user’s past preferences and identifying the sim- predictive reliability and validity [43, 44, 45]. For this study, the ilarities and di erences in preferences can be used to provide team used a three-point preference scale rather than the tradi- personalized spatial recommendations to individual users. tional seven-point thermal sensation scale, to limit subjectivity and make it easier for participants to frequently provide feed- back in field conditions. This saved participant’s e ort and time 4. Discussion in the field as well as helped channelize and organize data for the research team to work eciently. In general, the aim of The implementation of the platform resulted in various in- the comfort feedback prompts in the application (as shown in sights related to the work towards a space recommendation Figure 2d) was to seek answers to the following questions from system. Several lessons were learned from the related to de- participants regarding their perception of comfort: 1) Is their ployment, the methodology used, the use of the data for com- current condition comfortable or uncomfortable?; 2) Do they fort preference segmentation, and the foundation for automated desire any change?; and 3) If so, would they prefer warmer or means of space matching and allocation. cooler?. From a psychological point of view, the first question relates to the cognitive thermal state and the other two to the 4.1. Selecting a field-based experiment setup preferred thermal state based on previous studies [46, 47]. Thermal comfort research methodologies generally rely on two categories of implementation: laboratory-based methods 4.3. Identifying occupant comfort personality types (climate chambers) and field-based methods [39]. However, Researchers increasingly adapt data-driven methodologies while the internal design of a field-based study may not per- to address challenges of occupant satisfaction, environmental mit as rigorous statistical modeling and analysis as a carefully quality, and energy eciency in buildings today [48, 49, 50]. controlled climate chamber experiment, the field study serves This study uses data-driven methods to identify personalized a vital role in grounding the experimental findings in reality comfort profiles of users - clustering users into types based on by its relevance to building occupants going about their regular similar environmental preferences, as shown in Section 3. Such daily routines. This methodology provides external validity to results could be useful in multiple ways; for one, grouping users the experimental findings. It is also crucial for thermal comfort with similar environmental preferences could improve occupant practitioners interested in understanding the role of the disci- comfort, space, and energy eciency, as shown by other studies pline of environmental psychology in building comfort. [51]. Next, this method could also enhance the feedback given Past studies have classified the laboratory-based and field- to designers and operators about future building design features based methodologies as two fundamentally separate ap- and operating strategies to improve flexible workplace occupant proaches; a deterministic engineering approach versus a holis- satisfaction and performance [52]. In parallel, it is also easy to tic person-environment system architectural approach. The two see how the same methodology could be used to distinguish approaches di er based on the disciplines which conduct them spaces based on occupant comfort feedback data and IoT data - and their perception of the dynamic or static relationship be- to derive comfort profile types of spaces. tween occupant and buildings [39]. However, models derived from a deterministic approach, work well only within limited conditions, usually centrally controlled air-conditioned spaces. 4.4. Towards a space recommendation engine That scenario can be compared to holistic person-environment It is easy to see how the results from this study could be systems models that take into consideration a more extensive used to understand, and even predict, patterns and anomalies in range of conditions that building occupants may choose to make an occupant’s environmental preferences in flexible workspaces themselves comfortable, such as in naturally ventilated build- over time. Taking this a step further, learning from past comfort ings [40, 41]. Since one of the goals of this study is to un- preferences of occupants could be used to match them to spaces derstand the dynamic nature of occupant comfort in di erent with suitable environmental profiles with acceptable tempera- environmental and spatial contexts - the research team chose a ture, light, and noise levels on average. This process can be field-based experiment set up to provide higher ecological va- done in real-time using IoT data to test whether this leads to lidity to the findings compared to a lab experiment [42]. an increase in occupant satisfaction or performance in flexible workplaces compared to a baseline scenario. Such methods of 4.2. Using longitudinal data and a three-point preference scale suggesting or matching based on past preferences have been widely used in other industries such as media and social net- New technologies have made collection, processing, and working [53, 54, 55], but they are still a new concept for the analysis of large and complex data more manageable. Using built environment. these capabilities, this study utilizes QR codes, a mobile ap- plication, and a time-series database infrastructure for manage- Figure 6 illustrates an example of this potential matching ment of the data life cycle. It enables the processing and assess- paradigm as applied to thermal comfort. This figure shows the ment of a comparatively large comfort data set in a short time. distributions of dry bulb temperature for each of the Zones from Past studies in thermal comfort research have often referred to the case study as well as a heat map illustrating a subjectively- as five or seven-point scales. While valuable in some instances, selected Thermal Match Level based on the feedback illustrated recent work in this area has shown that user-friendly, simplified, in Figure 5. These results show how the segmentation created 9 Figure 6: Thermal comfort matching zone with preference type example. The box plots (left) illustrate the temperature ranges of each zone, while the heat map (right) shows the gradient of probability that each zone will be satisfactory for the comfort types segmented in Figure 5. Each of the zones has the potential to be a better or worse match for various comfort personality types - the higher the thermal match level, the more probability to meet the comfort preferences of each comfort personality type. This match level metric is the foundation for future work in the creation and testing of a recommendation system that automatically learns and suggests spaces to occupants. by the occupant data feedback histories can be used to match ever, it has stopped short of testing the ability to give the recom- them to spaces according to the match level. There is also the mendations and the reactions of users in the face of these sug- potential to interact with the building control systems to change gestions. An additional limitation for this study is that the sam- the conditions of the spaces to create more or less of a space ple size of participants is not large enough to make more gen- type based on the changing needs of building occupants. This eralizable characterizations of the comfort types and the vast paper illustrates that the collection of data and segmentation of range of behavior that occupants could exhibit. Also, the num- users is possible using the type of feedback data collected from ber and type of physical measurements in the spaces were not the test participants. Future work will investigate how this pro- exhaustive as phenomena such as radiant and space e ects were cess can be automated and metrics developed to show the suc- not measured. The next phase of the project is a spatial recom- cess of such matching in terms of reducing energy consumption mendation engine seeks to test the feature to suggest spaces to and improving thermal comfort. people in order for them to find available working spaces that match their immediate needs. This deployment could be framed in the same way that common platforms help people find a place 5. Conclusion to stay or find a ride. This platform design would then test tech- This paper describes the field-based implementation of a nologies such as desk recommendation (based on time duration, space allocation platform in six flexible working zones for oc- number of desks, noise levels and desk availability), and inte- cupant comfort data collection. Over a month, 25 participants gration with occupancy, intelligent power plugs, and building provided 1,182 environmental momentary assessment surveys management systems. Future deployments of the platform will of their thermal, visual, and aural comfort. This comprehen- focus on the data collection from a larger sample size, enabling sive data set provides exciting opportunities for interpreting and a much more generalizable characterization of comfort groups learning about occupant comfort behavior in built environments and measuring more physical parameters. through data-driven methods. By demonstrating how data can be utilized to group occupants into comfort profile types, this study can act as a potential stepping stone to related research ar- eas such as comfort profiling of spaces, occupant behavior anal- Conflict of Interest Statement ysis, and correlation identification between various spatiotem- poral variables in buildings. 5.1. Limitations The authors declare that the research was conducted in the This analysis has covered the deployment and collection of absence of any current commercial or financial relationships data from users of a matching-based flexible platform. How- that could be construed as a potential conflict of interest. 10 Author Contributions [12] D. Clements-Croome, Creative and productive workplaces: a review, In- telligent Buildings International 7 (2015) 164–183. [13] G. Kats, Green building costs and financial benefits, Massachusetts Tech- TS: platform design, infrastructure development, experimen- nology Collaborative Boston, MA, 2003. tal design, implementation and lead author of the paper; PJ: [14] A. Wilson, Making the case for green building, Environmental building conceptualization, author of the paper; CM: experimental de- news 14 (2005) 1–15. sign, data analysis, funding, project leadership, the correspond- [15] G. Brager, H. Zhang, E. Arens, Evolving opportunities for providing thermal comfort, Building Research & Information 43 (2015) 274–287. ing author of the paper. [16] W. O.Seppanen, Association of ventilation system type with sbs symp- toms in oce workers, Indoor Air 12 (2002) 98–112. [17] M. Frontczak, S. Schiavon, J. Goins, E. Arens, H. Zhang, P. Wargocki, Funding Quantitative relationships between occupant satisfaction and satisfaction aspects of indoor environmental quality and building design, Indoor Air The Ministry of Education (MOE) of Singapore 22 (2012) 119–131. [18] T. Cheung, S. Schiavon, T. Parkinson, P. Li, G. Brager, Analysis of the (R296000181133) and the National University of Singa- accuracy on pmv – ppd model using the ashrae global thermal comfort pore (R296000158646) provided support for the development database ii, Building and Environment 153 (2019) 205 – 217. and implementation of this research. [19] V. F. Licina, ˇ T. Cheung, H. Zhang, R. de Dear, T. Parkinson, E. Arens, C. Chun, S. Schiavon, M. Luo, G. Brager, P. Li, S. Kaam, M. A. Ade- bamowo, M. M. Andamon, F. Babich, C. Bouden, H. Bukovianska, Acknowledgments C. Candido, B. Cao, S. Carlucci, D. K. Cheong, J.-H. Choi, M. Cook, P. Cropper, M. Deuble, S. Heidari, M. Indraganti, Q. Jin, H. Kim, J. Kim, K. Konis, M. K. Singh, A. Kwok, R. Lamberts, D. Loveday, J. Langevin, The authors would like to acknowledge Prageeth Jayathissa S. Manu, C. Moosmann, F. Nicol, R. Ooka, N. A. Oseland, L. Pagliano, and Matias Quintana for their assistance in data collection and D. Petra ´s, ˇ R. Rawal, R. Romero, H. B. Rijal, C. Sekhar, M. Schweiker, processing, the NUS Department of Architecture and Building F. Tartarini, S. ichi Tanabe, K. W. Tham, D. Teli, J. Toftum, L. Toledo, for the use of their spaces, and Naor Biton’s e orts in the de- K. Tsuzuki, R. D. Vecchi, A. Wagner, Z. Wang, H. Wallbaum, L. Webb, L. Yang, Y. Zhu, Y. Zhai, Y. Zhang, X. Zhou, Development of the ashrae velopment of the platform. global thermal comfort database ii, Building and Environment 142 (2018) 502 – 512. [20] S. Ferrari, V. Zanotto, Adaptive comfort: Analysis and application of the Data Availability Statement main indices, Building and Environment 49 (2012) 25–32. [21] F. Nicol, F. Stevenson, Adaptive comfort in an unpredictable world, Segments of the raw data and analysis code used for this Building Research & Information 41 (2013) 255–258. study will be available in an open-access Github repository that [22] J. van Hoof, M. Mazej, J. L. M. Hensen, Thermal comfort: research and practice, Frontiers in Bioscience (Landmark Edition) 15 (2010) 765–788. includes further documentation . [23] T. Cheung, S. Schiavon, T. Parkinson, P. Li, G. Brager, Analysis of the accuracy on pmv–ppd model using the ashrae global thermal comfort database ii, Building and Environment (2019). References [24] V. F. Licina, ˇ T. Cheung, H. Zhang, R. De Dear, T. Parkinson, E. Arens, C. Chun, S. Schiavon, M. Luo, G. Brager, et al., Development of the [1] JLL, Occupancy benchmarking guide, 2018-2019, 2018. ashrae global thermal comfort database ii, Building and Environment [2] CBRE Workplace Strategy, Asia pacific major report - space utilisation: 142 (2018) 502–512. The next frontier september, 2015. [25] J. Kim, S. Schiavon, G. Brager, Personal comfort models–a new paradigm [3] Gensler, Innovation to support cres’ challenge, 2013. in thermal comfort for occupant-centric environmental control, Building [4] L. Engelen, J. Chau, S. Young, M. Mackey, D. Jeyapalan, A. Bauman, Is and Environment 132 (2018) 114–124. activity-based working impacting health, work performance and percep- [26] J. Kim, Y. Zhou, S. Schiavon, P. Raftery, G. Brager, Personal comfort tions? A systematic review, Building Research & Information 0 (2018) models: predicting individuals’ thermal preference using occupant heat- 1–12. ing and cooling behavior and machine learning, Building and Environ- [5] M. Weijs-Perree, ´ J. van de Koevering, R. Appel-Meulenbroek, T. Arentze, ment 129 (2018) 96–106. Analysing user preferences for co-working space characteristics, Build- [27] G. Brager, H. Zhang, E. Arens, Evolving opportunities for providing ing Research and Information 0 (2018) 1–15. thermal comfort, Building Research and Information 43 (2015) 274–287. [6] S. H. Ebrahim, Q. A. Ahmed, E. Gozzer, P. Schlagenhauf, Z. A. Memish, [28] M. Vesely, ´ P. Molenaar, M. Vos, R. Li, W. Zeiler, Personalized heating – Covid-19 and community mitigation strategies in a pandemic, BMJ 368 comparison of heaters and control modes, Building and Environment 112 (2020) m1066. (2017) 223 – 232. [7] JLL Research & Strategy, Covid-19 global real estate implications, 2020. [29] T. Sood, M. Quintana, J. Prageet, M. Abdelrahman, C. Miller, The sde4 [8] A. Hirst, Settlers, vagrants and mutual indi erence: Unintended conse- learning trail: Crowdsourcing occupant comfort feedback at a net-zero quences of hot-desking, Journal of Organizational Change Management energy building (2019) 06. 24 (2011) 767–788. [30] M. L. B. L. Y. H. A. G. Y. Z. Zhe Wang, Richard de Dear, Individual [9] J. G. Hoendervanger, I. De Been, N. W. Van Yperen, M. P. Mobach, di erence in thermal comfort: A literature review, Building and Environ- C. J. Albers, Flexibility in use: Switching behaviour and satisfaction ment 138 (2018) 181 – 193. in activity-based work environments, Journal of Corporate Real Estate 18 [31] OECD, Methodological considerations in the measurement of subjective (2016) 48–62. well-being, 2013. [10] L. Taskin, M. Parmentier, F. Stinglhamber, The dark side of oce de- [32] N. M. Bradburn, S. Sudman, B. Wansink, Asking questions: the definitive signs: towards de-humanization, New Technology, Work and Employ- guide to questionnaire design–for market research, political polls, and ment 34 (2019) 262–284. social and health questionnaires, John Wiley & Sons, 2004. [11] P. M. G. M. D. W. Milton, Donald K., Risk of sick leave associated with [33] M. S. Rahaman, H. Pare, J. Liono, F. D. Salim, Y. Ren, J. Chan, S. Kudo, outdoor air supply rate, humidification, and occupant complaints, Indoor T. Rawling, A. Sinickas, OccuSpace: Towards a robust occupancy pre- Air 10 (2000) 212–221. diction system for activity based workplace, in: 2019 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), pp. 415–418. https://github.com/buds-lab/spacematch-paper 11 [34] Y. Arakawa, Augmented workplace: Human-Sensor interaction for im- proving the work environment, in: Proceedings of the Augmented Hu- mans International Conference, number Article 35 in AHs ’20, Associa- tion for Computing Machinery, New York, NY, USA, 2020, pp. 1–2. [35] P. B. Cooper, K. Maraslis, T. Tryfonas, G. Oikonomou, An intelligent hot- desking model harnessing the power of occupancy sensing data, Facilities 35 (2017) 766–786. [36] S. M. Bae, K. H. Han, S. H. Choi, J.-H. Park, Development of seat rec- ommendation system for the collaboration in smart work center, Journal of Korean Institute of Industrial Engineers 40 (2014) 128–134. [37] D. S. Moskowitz, S. N. Young, Ecological momentary assessment: what it is and why it is a method of the future in clinical psychopharmacology, J. Psychiatry Neurosci. 31 (2006) 13–20. [38] L. Engelen, F. Held, Understanding the oce: Using ecological momen- tary assessment to measure activities, posture, social interactions, mood, and work performance at the workplace, Buildings 9 (2019) 54. [39] R. de Dear, Thermal comfort in practice, Indoor Air 14 (2004) 32–39. [40] R. de Dear, Global database of thermal comfort field experiments, ASHRAE Transactions 104 (1998) 1141–1152. [41] R. de Dear, G. S. Brager, Developing an adaptive model of thermal com- fort and preference (1998). [42] C. Andrade, Internal, external, and ecological validity in research design, conduct, and evaluation, Indian journal of psychological medicine 40 (2018) 498. [43] S. Dolnicar, B. Grun, User-friendliness of answer formats-an empirical comparison, Australasian Journal of Market & Social Research 15 (2007) [44] J. A. Krosnick, Questionnaire design, in: The Palgrave handbook of survey research, Springer, 2018, pp. 439–455. [45] S. Dolnicar, B. Grun, F. Leisch, J. Rossiter, Three good reasons not to use five and seven point likert items (2011). [46] M. Schweiker, X. Fuchs, S. Becker, M. Shukuya, M. Dovjak, M. Haw- ighorst, J. Kolarik, Challenging the assumptions for thermal sensation scales, Building Research & Information 45 (2017) 572–589. [47] K. Parsons, Human thermal environments: the e ects of hot, moderate, and cold environments on human health, comfort, and performance, CRC press, 2014. [48] M. Moezzi, J. Goins, Text mining for occupant perspectives on the phys- ical workplace, Building Research & Information 39 (2011) 169–182. [49] J. Goins, M. Moezzi, Linking occupant complaints to building perfor- mance, Building Research & Information 41 (2013) 361–372. [50] S. Altomonte, S. Schiavon, M. G. Kent, G. Brager, Indoor environmental quality and occupant satisfaction in green-certified buildings, Building Research & Information 47 (2019) 255–274. [51] N. Kamarulzaman, A. Saleh, S. Hashim, H. Hashim, A. Abdul-Ghani, An overview of the influence of physical oce environments towards em- ployee, Procedia Engineering 20 (2011) 262 – 268. 2nd International Building Control Conference. [52] M. Kwon, H. Remøy, M. van den Bogaard, Influential design factors on occupant satisfaction with indoor environment in workplaces, Building and Environment 157 (2019) 356 – 365. [53] X. Amatriain, J. Basilico, Recommender Systems in Industry: A Netflix Case Study, Springer US, Boston, MA, 2015. [54] E. Khazaei, A. Alimohammadi, An automatic user grouping model for a group recommender system in location-based social networks, ISPRS International Journal of Geo-Information 7 (2018) 67. [55] P. Resnick, H. R. Varian, Recommender systems, Commun. ACM 40 (1997) 56–58. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Statistics arXiv (Cornell University)

Spacematch: Using environmental preferences to match occupants to suitable activity-based workspaces

Sood, Tapeesh; Janssen, Patrick; Miller, Clayton
Statistics , Volume 2020 (2006) – Jun 17, 2020
Download PDF
12 pages

Loading next page...
 
/lp/arxiv-cornell-university/spacematch-using-environmental-preferences-to-match-occupants-to-P05XIZ0njC
ISSN
2297-3362
eISSN
ARCH-3347
DOI
10.3389/fbuil.2020.00113
Publisher site
See Article on Publisher Site

Abstract

The activity-based workspace (ABW) paradigm is becoming more popular in commercial oce spaces. In this strategy, occupants are given a choice of spaces to do their work and personal activities on a day-to-day basis. This paper shows the implementation and testing of the Spacematch platform that was designed to improve the allocation and management of ABW. An experiment was implemented to test the ability to characterize the preferences of occupants to match them with suitable environmentally- comfortable and spatially-ecient flexible workspaces. This approach connects occupants with a catalog of available work desks using a web-based mobile application and enables them to provide real-time environmental feedback. In this work, we tested the ability for this feedback data to be merged with indoor environmental values from Internet-of-Things (IoT) sensors to optimize space and energy use by grouping occupants with similar preferences. This paper outlines a case study implementation of this platform on two oce buildings. This deployment collected 1,182 responses from 25 field-based research participants over a 30- day study. From this initial data set, the results show that the ABW occupants can be segmented into specific types of users based on their accumulated preference data, and matching preferences can be derived to build a recommendation platform. Keywords: IoT - Internet of Things, Thermal comfort, Space utilisation, Flexible work arrangement, Activity-based workspaces 1. Introduction years have led them to rethink spatial density and utilization. In the past few years, rising corporate real estate (CRE) costs 1.1. The emergence of workplace flexibility and rapid changes in technology and nature of work have ren- dered inecient traditional modes of working where occupants In response to these challenges, new ways of working are are permanently designated a single work desk. Today, 37% evolving rapidly. These approaches aspire to simultaneously of all oce spaces are empty on any given workday [1], which balance operator’s cost and space saving demands with flexibil- equates to approximately 150 billion dollars annually in unused ity and comfort needs of employees through enabling occupant space globally [2]. These challenges are pushing building oper- mobility. Through most of these approaches, the occupant can ators to rethink occupant density and spatial utilization in work- work flexibly by choosing di erent spots within the workplace places. In a recent survey, while 95% of CRE professionals rather than being assigned a fixed desk as the one primary place believed that workplaces influence occupant productivity and of work. Once occupants are dynamic in the ways they use comfort, only one third measured that impact. Most others only space, it is easier to recapture underutilized spaces by operators considered traditional cost-based measures and metrics to quan- [2]. Workplace strategies of this kind are often referred to as tify workplace density and utilization [3]. Based on past stud- activity-based workspaces (ABW) or by other terms such as hot- ies, heavy reliance on cost-based measures generally results in desking, co-working, desk-sharing, flexible working, and oce operators taking away amenities that make occupants comfort- hoteling. Though each strategy varies slightly from the other, able and productive in cost-saving work environments. An ex- most promise benefits of improved spatial utilization and cost ample is the replacement of cubicles with benches to accommo- savings for operators while increasing overall comfort, choice, date a more significant headcount, removal of informal collabo- and control for occupants [4]. Equally, an extensive recent sur- ration spaces for more desks, or taking away of employee stor- vey of spaces utilizing one such strategy showed that the pri- age space all together [2]. Workplaces today face challenges mary motivations for occupants to work in such a workplace is as operators look to o set high rental costs by growing occu- because it allowed access to an inspiring work environment [5]. pant headcount within their existing footprint. Equally, rapid Understandably, an increase in the adoption of these concepts changes in technology and the nature of work in the past few can be seen in the growing co-working industry. Preprint submitted to Frontiers in Built Environment June 18, 2020 arXiv:2006.09570v1 [cs.CY] 17 Jun 2020 1.1.1. COVID-19 and its e ect on the future of work In response to these challenges, research in indoor occupant comfort has accelerated over the last twenty years. One of In early 2020, oce working habits of much of the world the many paradigm shifts has been the movement away from changed due to a global pandemic of a novel pathogen known traditional, physically-based deterministic models due to their as COVID-19. As there is no treatment or vaccine for this reported low accuracy (only 34%) across dozens of comfort virus, numerous community-driven mitigation strategies have studies in the past decades [18, 19]. Recent research e orts been deployed across the world. One of the most common is have progressed towards adaptive comfort models [20, 21, 22], the requirement for those that can work from home to do so which rely on human behavior. According to these models, dis- [6]. This exodus from oce spaces to the home has shown that comforting changes in the thermal environment are followed by such decentralization of oce work is possible, and desirable a behavioral change in people to restore comfort. Such actions in some situations. Going forward this forced push towards could include reducing individual activity levels or even open- home working will reinforce the need for corporate entities to ing a window. The main e ect of such models is to increase adopt more agile real-estate portfolios, with more flexibility and the range of conditions that designers can consider comfortable, a more distributed footprint to ease employees back into the for instance, naturally ventilated buildings in the tropics where workplace and cut commuting time to a single, large headquar- occupants have a higher degree of control over their thermal ters [7]. This digitization and rethinking of how people can environment. work could be a strong catalyst for the adoption of ABW-style Despite these advancements, even adaptive comfort models oce arrangements. follow a one-size-fits-all approach that ignores the personal as- 1.1.2. The limitations of ABW strategies pect of comfort; this is like expecting everyone to have sim- Despite the momentum towards ABW, there are a significant ilar preferences for food, music, or style; all subjective at- number of challenges that this strategy poses through its shift tributes of a person’s personality. Further work on adaptive in oce culture. A study focused on a sociological analysis of models is needed to identify comfort preferences on an individ- one approach showed a ”loss of everyday workspace ownership ual basis and not only based on the thermal conditions. Recent giving rise to practical and social tensions within the organiza- studies have shown that occupants exposed to the same condi- tion [8]”. Another study found lower than expected satisfaction tions could exhibit variations in environmental perception due with activity-based working environments due to rare switching to individual di erences in comfort preferences and personal- of di erent activity settings [9]. A recent study even found evi- ity [23, 24]. Researchers have addressed this through the de- dence of dehumanization as a result of ABW [10]. As organiza- velopment of personal comfort models that predict individual tions evolve, they are wary of ill-conceived applications which thermal comfort responses rather than the average response of a larger population [25] through leveraging machine learning may disrupt business and culture purely for cost savings of new techniques and the Internet of Things (IoT) technologies. This workplace strategies. These studies illustrate that there is much approach demonstrates a very high prediction accuracy, well improvement possible in the deployment of ABW strategies to beyond that of the traditional models [26]. help mitigate these downsides. 1.2. Connection to indoor environmental comfort in buildings 1.2.1. Building systems’ response to indoor environmental challenges In addition to space use allocation, indoor environmental comfort is at the forefront of building performance analysis. In addition to the work in comfort models, a large amount Occupant dissatisfaction with indoor environments has far- of research has also been done to improve the systems that re- reaching economic implications for workplaces. As people typ- spond to comfort needs. Various mechanical systems technolo- ically spend over 90% of their time indoors, indoor environment gies such as radiant systems and decentralized and personal- quality influences their comfort, performance, health, and well- ized ventilation attempt to address comfort problems. These being. Occupant dissatisfaction with indoor environments can contemporary innovations focus on the ability of a building to result in health impacts, absenteeism, and reduced productiv- adapt to its occupants by tracking them and modifying each per- ity [11]. Not only do enterprises today associate a majority son’s immediate personal climate to meet their individual needs of their costs (80-90%) to workers compensation and benefits [27]. [12, 13, 14] but as people typically spend more (about 90%) Personalized control system approaches have limitations as time indoors, the quality of the indoor environments influences the spatial resolution of most existing climate, lighting, and their comfort, performance and well-being at work. Continual noise control technologies does not have enough flexibility and occupant dissatisfaction with indoor environments can result responsiveness. Even more innovative, decentralized and per- in health impacts, absenteeism, and reduced productivity [11]. sonalized comfort systems are unable to create the response and A comparison of recent field studies from 467 air-conditioned resolution needed to practically create individualized comfort buildings containing 24,000 occupants showed between 30% zones for all occupants. Additionally, smaller and more decen- and 200% more cases of sick building syndrome symptoms than tralized systems create more maintenance tasks and complexity in the occupants of naturally ventilated buildings [15, 16]. An- within the building systems [28]. To meet these challenges of other survey in 2012, of 52,980 occupants in 351 oce build- decentralization, a balance between personalizing spaces and ings, found that 50% of the occupants were dissatisfied with maintaining the economies of scale that centralized systems their indoor environments [17]. provide could be achieved. Creating small zones of personal 2 comfort for all occupants should be used sparingly in particular there are three occupants with indoor comfort preferences: Per- space use types, while most other spaces can be conditioned to son A enjoys a dim, quiet, and moderately conditioned space, meet the needs of a subgroup of people. Person B is into a warmer and more active environment with a bit of noise and brightness, and Person C likes a cooler envi- ronment that is not dead silent. Perhaps these preferences are 1.2.2. Collecting human comfort feedback in buildings relatively consistent, or they could be dependent on the occu- Currently, researchers and building owners install a wide pant’s current activity or frame-of-mind. If these occupants’ range of IoT devices that measure various environmental condi- preferences are collected over time, the probability of nudg- tions such as light, noise, and particulates levels, in addition to ing these users to spaces in oces that match these specific the conventional temperature and humidity metrics. Although needs is higher. Also, there is the possibility of grouping peo- IoT sensors have become low cost and ubiquitous, the data from ple with similar preferences, which could be combined with these devices are often not utilized to their full potential. Com- systems control to create di erent zones with di erent types of fort models, even adaptive ones, only set thresholds to which comfort. Perhaps Person A, B, or C could be grouped with peo- sensor data points can be compared. The critical element in ple who have comfort personality type A, B, or C to improve putting these data in context is subjective and physiological both satisfaction and systems control. The goal of this study is feedback from people who inhabit the environment. Collecting to try to capture these tendencies using a web-based tool over a this type of data would empower more specialized and nuanced more extended period than a typical indoor comfort study. comfort models to be developed in a scalable way [29]. How- ever, the quantitative identification of individual di erences in comfort preferences and personality of people remains a sig- 1.3.1. Novelty of proposed approach nificant challenge in the field for researchers and practitioners [30]. There have been several recent e orts with the focus of using One way to tackle this situation is to use contemporary meth- technology to improve the ABW paradigm. One primary direc- ods of personal feedback collection such as structured surveys tion has been on the use of occupancy detection and prediction or interviews either online or oine, in-person or remote, which to characterize the use of ABW [33]. Additional work focuses would help increase the frequency and volume of building oc- on the use of human-sensor interaction to promote better deci- cupant feedback. However, such conventional methods have sions by occupants [34]. A previous study has used occupancy several shortcomings [31]. One major drawback of these meth- data to optimize the allocation of hot desk spaces using sim- ods is the lack of scalability. It is dicult to collect large sample ulation and occupancy sensors [35]. In the literature, there is data sets due to the administrative, financial, and other oper- a single case of designing a seat recommendation system that ational overheads associated with these approaches. Further- has initial e orts towards matching people to spaces that would more, other factors such as lack of knowledge (respondents best match their needs [36]. Therefore, the presented platform do not know the answer to a question, but answer it nonethe- and methodology is among the first examples of data collec- less), lack of motivation (respondents may not process ques- tion, characterization, and e orts towards a space recommen- tions fully) and failures in communication (survey questions dation system. The presented scope of work combines the use may be unclear or misunderstood) result in an increased risk of of clustering occupants based on their preferences gathered in biases and respondent heuristics in traditional survey responses a longitudinally-intensive method and characterizing a comfort [32]. As sensor adaptation in built environments continues to matching preference probability. These are techniques which grow, new technologies and modern data capabilities allow re- have not been found in the literature. searchers and practitioners to e ectively capture dynamic hu- man feedback e ectively. However, this approach also presents significant challenges in the collection, analysis, processing, 1.3.2. Organization of the paper and visualization of large data sets from building occupants. This study addresses each of the previously mentioned chal- There is a need for easy to use, scalable solutions that help iden- lenges: collecting larger volumes of personalized data useful tify and quantify occupant comfort preferences for operators as to occupants and operators, reducing the need for complex they move to ABW. and problematic personalized comfort systems, and impacting ABW by improving spatial utilization and occupant comfort. 1.3. Towards improving ABW by recommending the best loca- This recommendation system’s hypothesis is to test whether tion for an occupant based on their preferences certain groups of occupants can be segmented according to their This paper outlines a platform that improves indoor environ- comfort preferences and whether this segmentation is realistic mental satisfaction and ABW by allocating occupants to spaces in the context of an actual building case study. This paper shares that are the best match for their needs. We seek to test whether a the development and testing of the platform in the context of a longitudinally-intensive collection of indoor comfort data from case study implementation. In Section 2, we first illustrate an individual occupants can be used to assign each person to a cer- implementation with 25 research participants over a month in tain preference tendency type. The goal is to use these prefer- six flexible workspaces. Section 3 showcases the results from ence types to match that person to a space that could best meet this implementation and Section 4 discusses the interpretation their needs. For example, consider a simple situation in which of those results in the context of future work. 3 Figure 1: Overview of the platform user flow - the goal is to give flexible workspace occupants the ability to find spaces that meet their needs (1 and 2), give feedback about their comfort to develop a comfort personality type (3), and eventually provide train a model (4) comfort suggestions for subsequent uses (5) 2. Methodology implementation, as shown in Figure 3. The zones were split be- tween the two institutional buildings across three di erent lev- To test the user segmentation and space allocation hypothe- els. Each zone was strategically selected based on di erences sis, a progressive web application platform was developed for in location, floor level, number of desks, and zone accessibility. implementation in the SDE4 and SDE2 building on the cam- Further, di erences in window-to-wall ratios, zone orientation, pus of the National University of Singapore. The first step was and proximity to the nearby vehicular road and public areas of the development of the user flow of the platform. Figure 1 il- the institution lead to varying light and noise levels between lustrates the user in the case study who can find available open zones. workspaces, reserve them for use, and give comfort feedback to As shown in Figure 3, desks in each zone were arranged help train a model to predict which location would be best for in the proximity of fixed indoor environmental quality sensors them based on their past comfort feedback. To use the platform, measuring seven attributes in real-time: temperature, humid- the occupants can choose to use the work desk right away or re- ity, noise, light, carbon dioxide, volatile organic compounds, serve for later use. During use, the mobile application enables and presence. Each zone o ered flexible work desks in which occupants to quickly provide environmental comfort feedback each arrangement di ered slightly. In some zones, desks were for temperature, noise, and light variables, as shown in Figure arranged to promote collaboration; they were aligned so par- 2. For flexible workspace operators, it facilitates merging per- ticipants could face each other while working. In others, the sonalized environmental comfort data with other data streams arrangement enabled solitary, personal work. Each desk was such as indoor environmental quality, occupancy, and energy identified through a unique label containing the desk number, use data among others to optimize space, energy usage, and room name, and a QR code for connecting the desk to the occupant comfort by grouping users with similar preferences. progressive web application, as shown in Figure 4. During the study, participants were not assigned a specific desk and 2.1. Implementation were encouraged to alternate their desk usage between di er- ent zones rather than stick to a single zone. The experimental A case study was designed to test the platform in field condi- instruction were designed help to provide generality and eco- tions with 25 research participants over one month in April and logical validity to the collected data. May 2019. The participants were recruited from the current mix of students and sta , which were representative of regular users Each participant in the experiment used the interactive mo- of two selected institutional buildings. An ethics review was bile application to reserve and use work desks, as shown in Fig- submitted approved for the methodology of the study. A total ure 2. Participants could search for available work desks within of 36 desks in six di erent zones were identified for case study the two institutional buildings. Once they chose the building, 4 Figure 2: Overview of the application: (a) Find flexible working zones around campus, (b) Dashboard and information screen for more details regarding each zone, (c) Choose between options for use desk now or reserve desk for later, (d) Provide feedback for temperature, light and noise variables participants could progress to select the room and desk to use. At the start of the pilot study, a common onboarding session They were provided with more information about the room and for all participants was organized. During this session, goals, real-time indoor environmental quality through the Info and objectives, and the methodology for the study were discussed, Dashboard features of the application, as shown in Figure 2b. and the participants were onboarded to the interactive mobile The application provided options between using the desk right application. Using the platform, the research team demon- now or reserving it for later. To start using a desk, participants strated to the participants on how to find and book a desk and had to scan the desk QR code label using an in-built QR code to provide feedback during desk usage. Details such as partic- scanner in the application. A minimum of a two-hour time slot ipation schedule, timings, zone locations, and physical acces- for a work session was provided for each desk booking. Partic- sibility guidelines were also shared during this session. After ipants could also choose to extend their work sessions in two- this session, flexibility was provided for participants to use any hour multiples as per their requirements using the application. of the six zones between 8 am to 6 pm daily for the month dur- During use, the application prompted participants to provide ing the pilot study. Participants could also choose to participate environmental feedback for temperature, light, and noise levels in groups or individually based on their routine personal and through a three-point scale, as shown in Figure 2d. Prompts work preferences to ensure that there is no disturbance to the were configured such that the application nudged users to pro- typical flexible workspace environment. However, participants vide feedback at the start, finish, and once every half an hour of were encouraged to alternate their desk usage between di er- a typical two-hour work session. This miniature survey is a type ent zones and times of the day to provide variety, generality, of ecological momentary assessment, a method for longitudinal and ecological validity to the experimental findings. The test data collection pioneered in medicine and psychology [37] and participants gave individual feedback in a range of 40-100 total recently adapted more for environmental perception [38]. feedback points per person. 5 Figure 3: The six flexible working zones for case study implementation. The orange square in each zone layout represents the location of the indoor environmental quality sensors: (a) Layout and photo from Zone 1, (b) Zone 2, (c) Zone 3, (d) Zone 4, (e) Zone 5, (f) Zone 6, (g) Selection of the six zones across three floors between two separate institutional buildings The data from the users and fixed sensors were aggregated platform for data acquisition, storage, and error detection, as using a cloud-based, time-series database, which served as a shown in Figure 4. The combination of location-based user 6 Figure 4: Data exchange framework: (a) Desk QR code label, (b) Web-based mobile application, (c) IoT based indoor environmental quality sensors, (d) Time series data base comfort feedback and fixed environmental sensor data allowed grouping similar preference behavior. This e ort captures each clustering analysis for personalized comfort profiles of users. user’s behavior in their interaction with the system instead of These two data sources were merged through matching feed- the demographic, physiological, or environmental conditions back location (spatially localized through desk QR code label), variables that are typically addressed in environmental prefer- time of feedback collection (timestamp), and user ID (through ence studies. an in-built anonymous authentication method). 3.1. Discovering occupant personal comfort preference types 3. Results To cluster user preferences, unsupervised learning tech- niques were used to group the participants into cohorts with This section analyzes the environmental quality comfort similar feedback for temperature, light, and noise variables, as preferences for temperature, light, noise values from 25 re- shown in Figure 5. The analysis leads to identifying four dis- search participants in the pilot study. An unsupervised clus- tinct clusters based on di erences in preferences for tempera- tering technique is applied to segment participant comfort data, ture, light, and noise levels across participants. As shown in totaling 1,182 feedback points, into clusters based on similar Figure 5a, many participants are generally comfortable across behavior. This study focuses on a participant’s behavior based zones. However, some participants preferred cooler or warmer on their interaction with the system rather than on conventional environments. For participants in Hard to Predict cluster, variables in similar environmental preference studies, such as more extensive and more diverse data streams are needed to demographics, physiological, or environmental conditions. The understand their preferences better in the future. For visual emphasis is to apply an unsupervised clustering technique to comfort or light values related preferences, the clustering is the occupant data to segment the users who provide more than evenly spread between prefer dimmer and generally comfort- five feedback points into cohorts of similar behavior. This type able choices, as shown in Figure 5b. As can be observed, most of analysis focuses on the characterization of comfort prefer- participants would prefer a change in their light settings across ences in ways specific to each occupant, but generalizable by zones. As shown in Figure 5c, most participants were aurally 7 Figure 5: User comfort preference clustering based on: (a) thermal comfort feedback (degrees Celcius), (b) noise comfort feedback (dB), (c) light comfort feedback (lux) 8 comfortable with a few preferring a change in noise levels. Un- and easy-to-use measures can be employed without the loss of derstanding a user’s past preferences and identifying the sim- predictive reliability and validity [43, 44, 45]. For this study, the ilarities and di erences in preferences can be used to provide team used a three-point preference scale rather than the tradi- personalized spatial recommendations to individual users. tional seven-point thermal sensation scale, to limit subjectivity and make it easier for participants to frequently provide feed- back in field conditions. This saved participant’s e ort and time 4. Discussion in the field as well as helped channelize and organize data for the research team to work eciently. In general, the aim of The implementation of the platform resulted in various in- the comfort feedback prompts in the application (as shown in sights related to the work towards a space recommendation Figure 2d) was to seek answers to the following questions from system. Several lessons were learned from the related to de- participants regarding their perception of comfort: 1) Is their ployment, the methodology used, the use of the data for com- current condition comfortable or uncomfortable?; 2) Do they fort preference segmentation, and the foundation for automated desire any change?; and 3) If so, would they prefer warmer or means of space matching and allocation. cooler?. From a psychological point of view, the first question relates to the cognitive thermal state and the other two to the 4.1. Selecting a field-based experiment setup preferred thermal state based on previous studies [46, 47]. Thermal comfort research methodologies generally rely on two categories of implementation: laboratory-based methods 4.3. Identifying occupant comfort personality types (climate chambers) and field-based methods [39]. However, Researchers increasingly adapt data-driven methodologies while the internal design of a field-based study may not per- to address challenges of occupant satisfaction, environmental mit as rigorous statistical modeling and analysis as a carefully quality, and energy eciency in buildings today [48, 49, 50]. controlled climate chamber experiment, the field study serves This study uses data-driven methods to identify personalized a vital role in grounding the experimental findings in reality comfort profiles of users - clustering users into types based on by its relevance to building occupants going about their regular similar environmental preferences, as shown in Section 3. Such daily routines. This methodology provides external validity to results could be useful in multiple ways; for one, grouping users the experimental findings. It is also crucial for thermal comfort with similar environmental preferences could improve occupant practitioners interested in understanding the role of the disci- comfort, space, and energy eciency, as shown by other studies pline of environmental psychology in building comfort. [51]. Next, this method could also enhance the feedback given Past studies have classified the laboratory-based and field- to designers and operators about future building design features based methodologies as two fundamentally separate ap- and operating strategies to improve flexible workplace occupant proaches; a deterministic engineering approach versus a holis- satisfaction and performance [52]. In parallel, it is also easy to tic person-environment system architectural approach. The two see how the same methodology could be used to distinguish approaches di er based on the disciplines which conduct them spaces based on occupant comfort feedback data and IoT data - and their perception of the dynamic or static relationship be- to derive comfort profile types of spaces. tween occupant and buildings [39]. However, models derived from a deterministic approach, work well only within limited conditions, usually centrally controlled air-conditioned spaces. 4.4. Towards a space recommendation engine That scenario can be compared to holistic person-environment It is easy to see how the results from this study could be systems models that take into consideration a more extensive used to understand, and even predict, patterns and anomalies in range of conditions that building occupants may choose to make an occupant’s environmental preferences in flexible workspaces themselves comfortable, such as in naturally ventilated build- over time. Taking this a step further, learning from past comfort ings [40, 41]. Since one of the goals of this study is to un- preferences of occupants could be used to match them to spaces derstand the dynamic nature of occupant comfort in di erent with suitable environmental profiles with acceptable tempera- environmental and spatial contexts - the research team chose a ture, light, and noise levels on average. This process can be field-based experiment set up to provide higher ecological va- done in real-time using IoT data to test whether this leads to lidity to the findings compared to a lab experiment [42]. an increase in occupant satisfaction or performance in flexible workplaces compared to a baseline scenario. Such methods of 4.2. Using longitudinal data and a three-point preference scale suggesting or matching based on past preferences have been widely used in other industries such as media and social net- New technologies have made collection, processing, and working [53, 54, 55], but they are still a new concept for the analysis of large and complex data more manageable. Using built environment. these capabilities, this study utilizes QR codes, a mobile ap- plication, and a time-series database infrastructure for manage- Figure 6 illustrates an example of this potential matching ment of the data life cycle. It enables the processing and assess- paradigm as applied to thermal comfort. This figure shows the ment of a comparatively large comfort data set in a short time. distributions of dry bulb temperature for each of the Zones from Past studies in thermal comfort research have often referred to the case study as well as a heat map illustrating a subjectively- as five or seven-point scales. While valuable in some instances, selected Thermal Match Level based on the feedback illustrated recent work in this area has shown that user-friendly, simplified, in Figure 5. These results show how the segmentation created 9 Figure 6: Thermal comfort matching zone with preference type example. The box plots (left) illustrate the temperature ranges of each zone, while the heat map (right) shows the gradient of probability that each zone will be satisfactory for the comfort types segmented in Figure 5. Each of the zones has the potential to be a better or worse match for various comfort personality types - the higher the thermal match level, the more probability to meet the comfort preferences of each comfort personality type. This match level metric is the foundation for future work in the creation and testing of a recommendation system that automatically learns and suggests spaces to occupants. by the occupant data feedback histories can be used to match ever, it has stopped short of testing the ability to give the recom- them to spaces according to the match level. There is also the mendations and the reactions of users in the face of these sug- potential to interact with the building control systems to change gestions. An additional limitation for this study is that the sam- the conditions of the spaces to create more or less of a space ple size of participants is not large enough to make more gen- type based on the changing needs of building occupants. This eralizable characterizations of the comfort types and the vast paper illustrates that the collection of data and segmentation of range of behavior that occupants could exhibit. Also, the num- users is possible using the type of feedback data collected from ber and type of physical measurements in the spaces were not the test participants. Future work will investigate how this pro- exhaustive as phenomena such as radiant and space e ects were cess can be automated and metrics developed to show the suc- not measured. The next phase of the project is a spatial recom- cess of such matching in terms of reducing energy consumption mendation engine seeks to test the feature to suggest spaces to and improving thermal comfort. people in order for them to find available working spaces that match their immediate needs. This deployment could be framed in the same way that common platforms help people find a place 5. Conclusion to stay or find a ride. This platform design would then test tech- This paper describes the field-based implementation of a nologies such as desk recommendation (based on time duration, space allocation platform in six flexible working zones for oc- number of desks, noise levels and desk availability), and inte- cupant comfort data collection. Over a month, 25 participants gration with occupancy, intelligent power plugs, and building provided 1,182 environmental momentary assessment surveys management systems. Future deployments of the platform will of their thermal, visual, and aural comfort. This comprehen- focus on the data collection from a larger sample size, enabling sive data set provides exciting opportunities for interpreting and a much more generalizable characterization of comfort groups learning about occupant comfort behavior in built environments and measuring more physical parameters. through data-driven methods. By demonstrating how data can be utilized to group occupants into comfort profile types, this study can act as a potential stepping stone to related research ar- eas such as comfort profiling of spaces, occupant behavior anal- Conflict of Interest Statement ysis, and correlation identification between various spatiotem- poral variables in buildings. 5.1. Limitations The authors declare that the research was conducted in the This analysis has covered the deployment and collection of absence of any current commercial or financial relationships data from users of a matching-based flexible platform. How- that could be construed as a potential conflict of interest. 10 Author Contributions [12] D. Clements-Croome, Creative and productive workplaces: a review, In- telligent Buildings International 7 (2015) 164–183. [13] G. Kats, Green building costs and financial benefits, Massachusetts Tech- TS: platform design, infrastructure development, experimen- nology Collaborative Boston, MA, 2003. tal design, implementation and lead author of the paper; PJ: [14] A. Wilson, Making the case for green building, Environmental building conceptualization, author of the paper; CM: experimental de- news 14 (2005) 1–15. sign, data analysis, funding, project leadership, the correspond- [15] G. Brager, H. Zhang, E. Arens, Evolving opportunities for providing thermal comfort, Building Research & Information 43 (2015) 274–287. ing author of the paper. [16] W. O.Seppanen, Association of ventilation system type with sbs symp- toms in oce workers, Indoor Air 12 (2002) 98–112. [17] M. Frontczak, S. Schiavon, J. Goins, E. Arens, H. Zhang, P. Wargocki, Funding Quantitative relationships between occupant satisfaction and satisfaction aspects of indoor environmental quality and building design, Indoor Air The Ministry of Education (MOE) of Singapore 22 (2012) 119–131. [18] T. Cheung, S. Schiavon, T. Parkinson, P. Li, G. Brager, Analysis of the (R296000181133) and the National University of Singa- accuracy on pmv – ppd model using the ashrae global thermal comfort pore (R296000158646) provided support for the development database ii, Building and Environment 153 (2019) 205 – 217. and implementation of this research. [19] V. F. Licina, ˇ T. Cheung, H. Zhang, R. de Dear, T. Parkinson, E. Arens, C. Chun, S. Schiavon, M. Luo, G. Brager, P. Li, S. Kaam, M. A. Ade- bamowo, M. M. Andamon, F. Babich, C. Bouden, H. Bukovianska, Acknowledgments C. Candido, B. Cao, S. Carlucci, D. K. Cheong, J.-H. Choi, M. Cook, P. Cropper, M. Deuble, S. Heidari, M. Indraganti, Q. Jin, H. Kim, J. Kim, K. Konis, M. K. Singh, A. Kwok, R. Lamberts, D. Loveday, J. Langevin, The authors would like to acknowledge Prageeth Jayathissa S. Manu, C. Moosmann, F. Nicol, R. Ooka, N. A. Oseland, L. Pagliano, and Matias Quintana for their assistance in data collection and D. Petra ´s, ˇ R. Rawal, R. Romero, H. B. Rijal, C. Sekhar, M. Schweiker, processing, the NUS Department of Architecture and Building F. Tartarini, S. ichi Tanabe, K. W. Tham, D. Teli, J. Toftum, L. Toledo, for the use of their spaces, and Naor Biton’s e orts in the de- K. Tsuzuki, R. D. Vecchi, A. Wagner, Z. Wang, H. Wallbaum, L. Webb, L. Yang, Y. Zhu, Y. Zhai, Y. Zhang, X. Zhou, Development of the ashrae velopment of the platform. global thermal comfort database ii, Building and Environment 142 (2018) 502 – 512. [20] S. Ferrari, V. Zanotto, Adaptive comfort: Analysis and application of the Data Availability Statement main indices, Building and Environment 49 (2012) 25–32. [21] F. Nicol, F. Stevenson, Adaptive comfort in an unpredictable world, Segments of the raw data and analysis code used for this Building Research & Information 41 (2013) 255–258. study will be available in an open-access Github repository that [22] J. van Hoof, M. Mazej, J. L. M. Hensen, Thermal comfort: research and practice, Frontiers in Bioscience (Landmark Edition) 15 (2010) 765–788. includes further documentation . [23] T. Cheung, S. Schiavon, T. Parkinson, P. Li, G. Brager, Analysis of the accuracy on pmv–ppd model using the ashrae global thermal comfort database ii, Building and Environment (2019). References [24] V. F. Licina, ˇ T. Cheung, H. Zhang, R. De Dear, T. Parkinson, E. Arens, C. Chun, S. Schiavon, M. Luo, G. Brager, et al., Development of the [1] JLL, Occupancy benchmarking guide, 2018-2019, 2018. ashrae global thermal comfort database ii, Building and Environment [2] CBRE Workplace Strategy, Asia pacific major report - space utilisation: 142 (2018) 502–512. The next frontier september, 2015. [25] J. Kim, S. Schiavon, G. Brager, Personal comfort models–a new paradigm [3] Gensler, Innovation to support cres’ challenge, 2013. in thermal comfort for occupant-centric environmental control, Building [4] L. Engelen, J. Chau, S. Young, M. Mackey, D. Jeyapalan, A. Bauman, Is and Environment 132 (2018) 114–124. activity-based working impacting health, work performance and percep- [26] J. Kim, Y. Zhou, S. Schiavon, P. Raftery, G. Brager, Personal comfort tions? A systematic review, Building Research & Information 0 (2018) models: predicting individuals’ thermal preference using occupant heat- 1–12. ing and cooling behavior and machine learning, Building and Environ- [5] M. Weijs-Perree, ´ J. van de Koevering, R. Appel-Meulenbroek, T. Arentze, ment 129 (2018) 96–106. Analysing user preferences for co-working space characteristics, Build- [27] G. Brager, H. Zhang, E. Arens, Evolving opportunities for providing ing Research and Information 0 (2018) 1–15. thermal comfort, Building Research and Information 43 (2015) 274–287. [6] S. H. Ebrahim, Q. A. Ahmed, E. Gozzer, P. Schlagenhauf, Z. A. Memish, [28] M. Vesely, ´ P. Molenaar, M. Vos, R. Li, W. Zeiler, Personalized heating – Covid-19 and community mitigation strategies in a pandemic, BMJ 368 comparison of heaters and control modes, Building and Environment 112 (2020) m1066. (2017) 223 – 232. [7] JLL Research & Strategy, Covid-19 global real estate implications, 2020. [29] T. Sood, M. Quintana, J. Prageet, M. Abdelrahman, C. Miller, The sde4 [8] A. Hirst, Settlers, vagrants and mutual indi erence: Unintended conse- learning trail: Crowdsourcing occupant comfort feedback at a net-zero quences of hot-desking, Journal of Organizational Change Management energy building (2019) 06. 24 (2011) 767–788. [30] M. L. B. L. Y. H. A. G. Y. Z. Zhe Wang, Richard de Dear, Individual [9] J. G. Hoendervanger, I. De Been, N. W. Van Yperen, M. P. Mobach, di erence in thermal comfort: A literature review, Building and Environ- C. J. Albers, Flexibility in use: Switching behaviour and satisfaction ment 138 (2018) 181 – 193. in activity-based work environments, Journal of Corporate Real Estate 18 [31] OECD, Methodological considerations in the measurement of subjective (2016) 48–62. well-being, 2013. [10] L. Taskin, M. Parmentier, F. Stinglhamber, The dark side of oce de- [32] N. M. Bradburn, S. Sudman, B. Wansink, Asking questions: the definitive signs: towards de-humanization, New Technology, Work and Employ- guide to questionnaire design–for market research, political polls, and ment 34 (2019) 262–284. social and health questionnaires, John Wiley & Sons, 2004. [11] P. M. G. M. D. W. Milton, Donald K., Risk of sick leave associated with [33] M. S. Rahaman, H. Pare, J. Liono, F. D. Salim, Y. Ren, J. Chan, S. Kudo, outdoor air supply rate, humidification, and occupant complaints, Indoor T. Rawling, A. Sinickas, OccuSpace: Towards a robust occupancy pre- Air 10 (2000) 212–221. diction system for activity based workplace, in: 2019 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), pp. 415–418. https://github.com/buds-lab/spacematch-paper 11 [34] Y. Arakawa, Augmented workplace: Human-Sensor interaction for im- proving the work environment, in: Proceedings of the Augmented Hu- mans International Conference, number Article 35 in AHs ’20, Associa- tion for Computing Machinery, New York, NY, USA, 2020, pp. 1–2. [35] P. B. Cooper, K. Maraslis, T. Tryfonas, G. Oikonomou, An intelligent hot- desking model harnessing the power of occupancy sensing data, Facilities 35 (2017) 766–786. [36] S. M. Bae, K. H. Han, S. H. Choi, J.-H. Park, Development of seat rec- ommendation system for the collaboration in smart work center, Journal of Korean Institute of Industrial Engineers 40 (2014) 128–134. [37] D. S. Moskowitz, S. N. Young, Ecological momentary assessment: what it is and why it is a method of the future in clinical psychopharmacology, J. Psychiatry Neurosci. 31 (2006) 13–20. [38] L. Engelen, F. Held, Understanding the oce: Using ecological momen- tary assessment to measure activities, posture, social interactions, mood, and work performance at the workplace, Buildings 9 (2019) 54. [39] R. de Dear, Thermal comfort in practice, Indoor Air 14 (2004) 32–39. [40] R. de Dear, Global database of thermal comfort field experiments, ASHRAE Transactions 104 (1998) 1141–1152. [41] R. de Dear, G. S. Brager, Developing an adaptive model of thermal com- fort and preference (1998). [42] C. Andrade, Internal, external, and ecological validity in research design, conduct, and evaluation, Indian journal of psychological medicine 40 (2018) 498. [43] S. Dolnicar, B. Grun, User-friendliness of answer formats-an empirical comparison, Australasian Journal of Market & Social Research 15 (2007) [44] J. A. Krosnick, Questionnaire design, in: The Palgrave handbook of survey research, Springer, 2018, pp. 439–455. [45] S. Dolnicar, B. Grun, F. Leisch, J. Rossiter, Three good reasons not to use five and seven point likert items (2011). [46] M. Schweiker, X. Fuchs, S. Becker, M. Shukuya, M. Dovjak, M. Haw- ighorst, J. Kolarik, Challenging the assumptions for thermal sensation scales, Building Research & Information 45 (2017) 572–589. [47] K. Parsons, Human thermal environments: the e ects of hot, moderate, and cold environments on human health, comfort, and performance, CRC press, 2014. [48] M. Moezzi, J. Goins, Text mining for occupant perspectives on the phys- ical workplace, Building Research & Information 39 (2011) 169–182. [49] J. Goins, M. Moezzi, Linking occupant complaints to building perfor- mance, Building Research & Information 41 (2013) 361–372. [50] S. Altomonte, S. Schiavon, M. G. Kent, G. Brager, Indoor environmental quality and occupant satisfaction in green-certified buildings, Building Research & Information 47 (2019) 255–274. [51] N. Kamarulzaman, A. Saleh, S. Hashim, H. Hashim, A. Abdul-Ghani, An overview of the influence of physical oce environments towards em- ployee, Procedia Engineering 20 (2011) 262 – 268. 2nd International Building Control Conference. [52] M. Kwon, H. Remøy, M. van den Bogaard, Influential design factors on occupant satisfaction with indoor environment in workplaces, Building and Environment 157 (2019) 356 – 365. [53] X. Amatriain, J. Basilico, Recommender Systems in Industry: A Netflix Case Study, Springer US, Boston, MA, 2015. [54] E. Khazaei, A. Alimohammadi, An automatic user grouping model for a group recommender system in location-based social networks, ISPRS International Journal of Geo-Information 7 (2018) 67. [55] P. Resnick, H. R. Varian, Recommender systems, Commun. ACM 40 (1997) 56–58.

Journal

Statistics – arXiv (Cornell University)

Published: Jun 17, 2020

References

  • Indoor environmental quality and occupant satisfaction in green-certified buildings
    Altomonte