Purpose: This paper seeks to pilot test a novel way to collect freight and service activity data and analyze the collected data in the metropolitan area of Medellin, Colombia. Methods: This research collects data using a multi-layer and multi-actor approach that includes surveys to receivers, suppliers, carriers, and truck drivers. The data are used by the authors to describe the overall freight patterns in the area of study and to show lessons learned. Results: The data collection resulted in 2947 establishments (4.4% of the total establishments in the city), a cordon survey of 2950 commercial vehicles (17% of the total vehicle volume) accessing the urban area, and carrier interviews to ten companies and 130 truck drivers. The results indicate that a total of 33,274 metric tons/day enter the study area, 35,240 tons/day leave the area; while 7000 tons/day are distributed in the study area. In terms of freight trips, 6600 trips/day enter the study zone and 6600 trips/day leave it. Conclusions: The data collection effort enabled the analyses of freight generation patterns. The freight surveys used in the study complement each other, and provided a good depiction of the freight movements in urban areas. It was found that in the Medellin Metropolitan Area, freight-intensive sector establishments generates, on average, significantly more cargo (freight attraction plus production) than the service-intensive sectors. The analyses of the surveys allow the decision makers to understand the nature of the cargo and the generation patterns in different type of establishments. This characterization of the freight patterns is vital for the forecasting of the behavior of the cargo and it is the main input to perform freight demand modeling for city planning, especially for developing countries, where there are too many budget constraints. Keywords: Urban freight, Cordon survey, Establishments’ survey, Urban freight patterns 1 Introduction Freight activity is difficult to understand because there The movement of goods is an essential factor in the are many actors involved (e.g., shippers, carriers, distrib- economy of cities, regions, and countries. In recent years utors, freight forwarders, receivers) involved in multiple the quest for sustainability has motivated policy makers industry segments and not much information is available to find strategies that reduce the externalities produced about how they operate. As a result, characterizing (e.g., congestion, pollution, noise) while protecting and freight patterns require a multi-layer and multi-actor ap- enhancing its economic benefits. It is crucial to under- proach. Goods movements can be broadly divided in stand the patterns of goods movements to be able to for- three clusters: (1) the flows of commodities resulting mulate alternatives that advance sustainability and from economic transactions that arrive through inter- economic progress. national gateways (e.g., ports, airports) and are then transported to manufacturing districts or to wholesalers for distribution; (2) the flows of commodities that are * Correspondence: email@example.com consumed by manufacturers, transformed, and then Department of Civil Engineering, Universidad Nacional de Colombia at shipped again to local or international destinations; and Medellin, Calle 65 # 78-28, M1-110 Medellin, Colombia Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 2 of 11 (3) the flows of commodities that are distributed for final are difficult to conduct because of the confidentiality of consumption (e.g., supplies for retail stores). These three the data; and (3) carrier surveys that quantify freight pat- types of flows converge in urban environments. While terns in terms of commercial vehicle flows in the trans- most of the attention has traditionally focused on the first portation network . Considering the information that and the second clusters because they move the largest can be captured from different type of surveys, profes- amounts of cargo, the bulk of freight trips—and their sionals in charge of estimating freight demand in urban negative impacts—are produced by urban deliveries, that regions must identify the commercial characteristics of are actually the focus of the “City Logistics” initiatives. the city in order to select the best type of freight surveys Complicating things there is a lack of data about to be conducted. It is important to note that typically freight demand, especially in developing countries. The one survey type cannot capture all the freight informa- lack of data is the result of multiple factors: the difficulty tion in a region. For example, origin-destination sur- of collecting the data, lack of awareness about the im- veys—which are the most common method to obtain portance of proactive public-sector involvement in information of truck travel patterns in a city—do not freight transportation management, the cost of collecting allow to draw a complete picture of all the freight pat- the data, among other reasons. terns in a city because they miss freight data (e.g., traffic The recent data collection in Medellin (the data were counts, freight generation) to represent the freight traffic collected in 2011–2012 by the Universidad Nacional de flows on the transportation network. Colombia at Medellin—that is the focus of this paper—is Allen, et al.  presented a complete revision of 162 an interesting case because it is a comprehensive effort studies of urban freight that collected urban freight data to characterize a city’s freight activity patterns. from the 1960’s to 2008in18 countries.Only 28studies The freight study reported in this paper is unique be- were made in the period 1960–1989. The rest (134) were cause it pilot test the freight data collection approach made from 1990 to 2008 showing an increment of the outlined by . This innovative approach pieces together practice in recent years, especially in Europe. This increase different surveys to carriers, receivers, manufacturers, in the number of freight studies could be due to new tech- and others; and combines data collection and modeling niques in collecting data and interest of the public sector to obtain fine-level detail of freight activity. Additional in the freight movements. However, it is important to note details are presented in section 5. that many freight data collection efforts are not published To the best of the authors’ knowledge, this is the first in journals or in English. Most of them are reports written freight survey done in the Metropolitan Area of Medellin; in the language of the region/country that sometimes are for that reason, the results obtained are essential to sup- not disseminated or simply archived. port freight policy making in the Medellin Metropolitan This paper describes the process followed and presents Area. The lessons learned in this effort are relevant to the results obtained from a multilayer data collection ap- countries all over the world where there are limited proaches based on Holguín-Veras and Jaller . This ap- resources for freight data collection. proach includes a cordon survey, an establishment-based This paper has five additional sections to this intro- survey, and a carrier survey. This gathering of freight duction. Section 2 gives a background summary of data is essential to city planning due to the economic freight studies. Section 3 presents the description of the implications of it. The study presented here provides an case study characterizing the cargo in Medellin. Section example of how the freight information collected in 4 shows the data collection conducted. Section 5 Medellin was used to obtain freight patterns of this de- analyses the results emphasizing on freight generation veloping city, and eventually the process could be ap- patterns. Finally, Section 6 summarizes the main plied in similar urban areas. findings and conclusions. 3 Area description 2 Background summary on freight studies The Medellin Metropolitan Area (MMA) includes the To understand the system as a whole, relevant data need city of Medellin (the second largest city in Colombia) to be collected. Usually this information is obtained and nine other urban areas, with approximately 3.5 mil- through surveys that are carried out at different places. lion inhabitants and 67,800 commercial establishments The movement of cargo in a region can be captured ba- (19.4 establishments for every 1000 inhabitants) in 2012 sically from three types of freight surveys that comple- . Medellin is an important industrial city in Colombia ment each other (1) establishment surveys that target due to its manufacturing activity and its contribution to the producers and consumers of goods to estimate GDP, the largest one after Bogota, the capital . The freight production and attraction inside the urban area; MMA has an international airport to plays an important (2) distributor surveys that capture information on role in the import and export of goods, especially goods movements through intermediaries but sometimes flowers. Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 3 of 11 The Northern counties represent 15% of the popula- 4 Data collection tion but account for less than 4% of the establishments This section describes the methodology conducted to in the study area; a likely consequence of the low income obtain the freight data in the study area. The first sub- nature of this part of the metropolitan area. Conversely, section explains different freight surveys. The second the Southern counties, with a similar share of popula- subsection summarizes the freight surveys conducted in tion, concentrate 10% of the establishments. This reflects the study. their relatively high income and the presence of manu- facturing in this part of the MMA. The urban core, 4.1 Brief description of overall approach / methodology Medellin, gathers 70% of the population and about 85% This paper is based on the work done by Holguín-Veras, of the establishments. Needless to say, the urban core et al. . The authors identified the main freight de- plays a major role in terms of commercial activity, suf- mand modeling techniques, determined their main data fering the largest freight–related traffic impacts, even in requirements, proposed a data collection methodology the main road corridor that traverses the city. combining multiple methods, measuring different units The zoning system used in the study consists of 456 and targeting multiple agents to obtain a comprehensive transportation analysis zones (TAZs). Those TAZs are view of freight activity. In doing so, the authors based on previously studied areas/regions of the city  proposed a data collection framework that includes that facilitate the analyses of the results and comparison establishment-based surveys (receiver, shipper, and car- with previous transportation studies. The geographical rier), trip intercept-based surveys, cordon surveys, distribution of the establishments in the city of vehicle-based surveys, and tour-based surveys. They sug- Medellin—where the freight surveys took place—is gested that the data collected should meet the expect- shown in Fig. 1. The blue color indicates less presence of ation of the planning agency that is going to use them, establishments while the red color indicates the opposite. should be cost-efficient, should be oriented to fill data The circles in the figure show the Central Business gaps, and become the basis for freight demand synthesis. District (CBD) and El Poblado—a dense commercial area In other words: “the best approach will be the combin- in the south of Medellin where the highest income popu- ation of data collection procedures (that best fits the lation live, which is emerging as a secondary business dis- needs and constraints of the participating agencies)” . trict—. The CBD and El Poblado concentrate more than The main intent of the freight data collection effort 60% of the commercial establishments in the metropolitan was to characterize the freight activity that takes place in area, which generate large freight and trips movements in all stages of the supply chains in the MMA. Some of the the area. The areas with no establishments (blank spaces) key data to be collected includes: 1) origin and destin- in the map correspond to hills, rivers, and the like. ation of the trip; 2) commodities being transported Fig. 1 Zoning system and location of commercial establishments in Medellin Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 4 of 11 (quantity, unit, weight, and industrial sector); 3) com- supply chain management between establishments, mercial vehicle being used (vehicle type, vehicle config- their suppliers and freight transport operators. The uration, number of axels, number of stops per tour, survey should be done with the help of the private capacity, percentage of cargo transported, and load fac- sector officials of the city to facilitate the process. tor); 4) vehicle delivery/collection trips at establishments; 5) goods flows to/from establishments; 6) loading/ 4.2 Data collection unloading activity of goods vehicles; 7) parking activity The data collection procedure used by the authors was of trucks; 8) trip details and patterns of goods vehicles; based on the following components: 7) conveyance of goods between vehicles and establish- ments; 10) origin location of goods flow/vehicle trip to Cordon survey: The survey was conducted between establishment; 11) supply chain management between October and November of 2011 to 2950 commercial establishments, their suppliers and freight transport op- vehicles (17% of the incoming flow) in 11 tollbooths in erators; and 12) Goods order (food/pharmacy) from the boundaries of Medellin to capture the flows stores nearby. entering and exiting the city. The surveys were Some of the key mechanisms to collect freight data conducted, for security and practical purposes from include: 6 AM to 7 PM and with the help of the highway police. Establishment survey: The establishments were geo- Cordon survey: During the week days, it is common located and classified in 20 industry sectors using that about 50% of the traffic entering the city is the International Industrial Uniform Code (IIUC) comprised of large trucks, thus the importance of . Thirteen sectors were classified as freight- measuring the entering flows though cordon surveys intensive sectors (e.g., mineral; construction; manu- . The questionnaire should contain questions about facturing; accommodation and food, transportation; origin and destination of the trip, commodities being wholesale trade; retail trade; food) while the other transported (quantity, unit, weight, and industrial seven are service-oriented (e.g., finance, insurance sector) and the commercial vehicle being used and real estate; service industries; public administra- (vehicle type, vehicle configuration, number of axels, tion). The geo-location of the establishments in the number of stops per tour, capacity, percentage of study area made possible the design of a sample that cargo transported, and load factor). respects the geographical breakdown of establish- Establishment survey: The establishments need to be ments with about 13% at north, 13% at south, and geo-located and classified in industry sectors using 74% in Medellin. The sample was randomly drawn an appropriate classification system differentiating across the ten municipalities of the region. Only 1% between freight-intensive sectors (e.g., mineral; con- of the establishments listed were not considered in struction; manufacturing; accommodation and food, the sample because they are located in rural areas transportation; wholesale trade; retail trade; food) not covered by the survey. Similarly to the geograph- and service-intensive sectors (e.g., finance, insurance ical distribution, the sample was designed to repro- and real estate; service industries; public administra- duce the distribution of industry segments with the tion). The data need to be collected through surveys. exception of the large establishments (more than A set of expansion factors, relating the sample to the 200 employees), due to lack of data for this type of universe for each industry code, needs to be applied establishements. Most of the IIUCs were considered to the survey results to obtain the overall picture of in the sample. The only exceptions were agriculture, internal freight activity. The questionnaire should livestock, fishing (usually not present in urban contain questions about vehicle delivery/collection areas), other housing service activities, and other trips at establishments; goods flows to/from estab- services. The data were collected through surveys in lishments; loading/unloading activity of goods vehi- 4 months (between February and May of 2012) with cles; and parking activity of service vehicles. a sample of 2947 establishments (4.4% of the total). Carrier interviews: to complement the establishment About 2% had major errors and were excluded from and cordon surveys, a carrier survey should be the expansion procedure. A set of expansion factors, conducted to companies and truck drivers at the relating the sample to the universe for each IIUC, urban area to validate some practices of behavioral were applied to the survey results to obtain the over- aspects of the urban deliveries and logistics practices all picture of internal freight activity. for different industry sectors. The questionnaire Carrier interviews: to complement the establishment should contain questions about trip details and and cordon surveys, a carrier survey was applied to patterns of goods and service vehicles; conveyance ten companies and 130 truck drivers at the urban of goods between vehicles and establishments; and area to validate some practices of behavioral aspects Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 5 of 11 of the urban deliveries and logistics practices for per person in the Medellin Metropolitan Area is about different industry sectors. The survey was done with 25 kg/day (total cargo divided by population), without the help of the police, which collaborated in this taking into account the freight passing through the re- process in 2012. gion. This value is an interesting indicator related to the intensity of freight activity and consumption patterns in Table 1 summarizes the freight data used for the study. the region. The amount of cargo consumed by person Once the data were collected and cleaned, the authors in a city is related to the proportion of freight activity proceeded with their descriptive analysis that is pre- (population and number establishments) in the regions sented in the next section. and to the purchasing power of people. Table 2 shows the commodities transported by vehicle 5 Results type. As shown, in terms of daily trips, 17,550 commer- This section presents the descriptive analyses of the data. cial vehicles enter, leave, or pass through the metropol- The first part of the section describes the results for the itan area, and 22% of the total trips are empty. This cordon surveys, while the second presents the results for proportion of empty trips is a typical pattern in the establishments, and carriers, surveys. The results Colombia and other countries [6–9]. Moreover, in terms should be analyzed as a global finding because each of freight trips, the results show that only 10% of daily freight survey complements the rest. trips are external trips, the remaining 90% are internal. It is noteworthy that 21% of the cargo is transported in 5.1 Cordon surveys large trucks with large shipment sizes. This reflects the The results from the cordon survey provide a panoramic fact that the majority of the cargo inside the city is being view of freight flows of the MMA. The results show that transported in small trucks, autos, and motorcycles 33,274 tons enter the metropolitan area every day; (79%). That produce 90% of the daily freight trips with- 35,240 tons leave it, and additionally, 27,653 are external out doubt the large traffic or small trucks create negative to external flows. An additional 2400 tons of trash as externalities such as congestion, emissions, and noise. well as 8000 tons of construction materials waste The commodities transported that have the largest mar- (debris) are produced and transported every day within ket share in the MMA are: Manufacturing (22%), Food the metropolitan area. The external-internal and (21%); Parcels (13%), Construction (11%), Chemicals internal-external flows are similar (30% of the total each) (8%), and Animals (3%). The results also indicate that and 85% of the cargo is transported by trucks. Inside the approximately 4% of the vehicles entering the region region, 15% of the cargo is composed by the flows be- were light vehicles (autos, pick-ups, and motorcycles), tween small establishments, debris, and trash. 11% small trucks and 71% medium trucks and 14% large The results of the cordon survey show that consider- trucks. These shares of vehicle types are different from ing all the cargo in the city, the amount of freight moved the ones in the city where small vehicles are the most common. The vast majority of the vehicles in the cordon Table 1 Freight data used for the study survey are medium-large trucks (85%) which transport Data collection Observations Variables large amounts of cargo in and out the city. In terms of trip chains, the surveys inquired about the Cordon 2950 commercial Origin and destination of the trip survey vehicles number of stops per tour by commodity type. As shown Commodities being transported in Holguín-Veras , the number of tours depends on Commercial vehicle being used the characteristics of the country, city, type of truck, the Establishment 2947 establishments Vehicle delivery/collection number of trip chains, type of carrier, service time, and survey trips at establishments commodity transported. Goods flows to/from In terms of commodity type, the authors found that the establishments average number of stops vary between 1 (e.g., magazines, Loading/unloading activity of coffee, leather) and 10 (e.g., lubricants, beverages, goods vehicles aluminum) with an average of 2.6 stops per tour as shown Parking activity of trucks in Fig. 2. Moreover, in general 66% of the trucks make 1 Carrier 10 companies Trip details and patterns of stop (only one destination), 31% make between 2 and 10 interview and 130 truck goods vehicles drivers stops, and 3% make more than 10 stops. As expected, the Conveyance of goods between frequency reduces as the number of stops increases. vehicles and establishments Supply chain management 5.2 Establishment and carrier surveys between establishments, their suppliers and freight transport The results from the establishment survey provides a operators panoramic view of freight production and attraction in Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 6 of 11 Table 2 Commodities transported by vehicle type Vehicle/commodity Food Animals Construction Parcels Manufacturing Chemical Empty Total Small vehicles 2943 427 1416 2010 3338 677 3140 13,950 Large vehicles 749 60 523 238 504 751 776 3601 Total 3693 487 1939 2248 3841 1428 3917 17,551 the metropolitan area of Medellin. Although the cordon was variable even in the same industry sector. These long survey showed that about 33,000 tons enter and leave the parking times, combined with the lack of appropriate area of study every day, the establishment survey shows loading/unloading facilities, have negative effects in terms that about 7000 tons are transported between small and of street blockages and the resulting congestion. medium size establishments; and 1000 tons have a destin- Table 3 shows the breakdown of trips and tonnage by ation outside the region, then the freight movement inside industry sector. After applying the expansion factors to the metropolitan area reaches 6000 ton/day. In terms of the samples, the total number of trips and tonnage were trips, the bulk is generated by urban deliveries. Small es- computed by industry sector. tablishments produce and attract 94,000 trips which is A breakdown by industry sector reveals that most of the about 6 times the near 17,000 truck trips captured in the observations are in the wholesale trade sector (34%), retail external cordon survey. In contrast, the tonnage entering (28%), and accommodation and food (16%). In terms of and leaving the region is near 12 times the tonnage pro- business size, most of the establishments in the sample are duced and attracted inside the MMA. Considering the small to medium-sized, i.e., less than 50 employees (90%). number of trips and cargo, we noticed that large establish- Furthermore, 90% of the establishments in Medellin have ments (establishments with more than 200 employees) is less than 10 employees, and 8% have between 11 and 50 not the main source of the freight traffic, but the large employees. All these establishments are micro and small number of trips of small vehicles are produced inside this business with an average of 3.2 employees. area to distribute the cargo. The results indicate that the large freight generators The analysis shows that 71% of the establishments do (e.g., large buildings, universities, hospitals) concentrate not have storage space. This creates the imperative need 26% of the cargo with only 4.6% of the trips in the city. for frequent deliveries. It was also found that 45% of the The manufacturing, retail, and services sectors concen- establishments open between 7:00 and 8:00 a.m., and 40% trate about 80% of the deliveries and 53% of tonnage. of the establishments close between 5:00 and 6:00 p.m. Every delivery of these three sectors weights on average 87% of the cargo by tonnage is moved by trucks, pick-ups less than 50 kg. If the large generators are included the and autos, the remaining 13% if transported by motor- number of deliveries reaches about 85% and the tonnage cycle. Data show that 70% of the trucks park in the street around 80%. In terms of the delivery size, the average for loading/unloading (in 20% of the cases they reported looks small with 64 kg per delivery. It could be due to to park illegally though the real number could be higher). the high presence of the retail sector in the city logistics. Of great interest to this paper, the average time for load- However, large generators receive the heaviest deliveries ing/unloading vehicles is 18.5 min. This is a general aver- with 361 kg, and other sectors like wholesale trade age for all sectors because the loading/unloading activity receive 110 kg per delivery on average. Fig. 2 Average number of stops per commodity type Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 7 of 11 Table 3 Breakdown of trips and tonnage by industry classification Industry classification Trips/day % of trips Tons/day % of tonnage Kg/trip Retail trade 28,018 29.7% 1200 19.8% 42.8 Manufacturing 24,812 26.3% 1317 21.7% 53.1 Wholesale trade 6920 7.3% 763 12.6% 110.3 Construction 4170 4.4% 300 5.0% 71.9 For hire carriage 2361 2.5% 59 1.0% 25.0 Services 21,665 23.0% 695 11.5% 32.1 Utilities/communication 183 0.2% 7 0.1% 38.3 Public administration 12 0.0% 3 0.0% 250.0 Large freight generators 4364 4.6% 1577 26.0% 361.4 Agriculture 0 0.0% 0 0.0% 0.0 Mining 1528 1.6% 109 1.8% 71.3 Unclassified 151 0.2% 30 0.5% 198.7 Total 94,184 100.0% 6060 100.0% 64.3 Another interesting finding from the study was the ve- establishments), while for the latter the main source of hicle classes used to transport the freight within the revenue are service activities (e.g., public services, financial metropolitan area. Table 4 shows the daily breakdown of services). The overall average for employees per establish- trips and tonnage by vehicle class for the sample. ment is 9.1 employees. The average employment for FIS As showninTable 4, more than two thirds of freight establishments is 8.8, while the average for non-FIS ones inside the metropolitan area are transported in trucks. is 10.2. The transport/warehousing sector has the highest Another 33% of freight is transported in smaller vehi- average of employees per establishment (39.7), followed cles: 11% transported by pickups and vans, about 9% by by service-oriented sectors education (17.13), health and cars and about 14% by motorcycles and bicycles. As social services (17.13), and financial services (17.01). shown, bicycles and motorcycles produce a large In terms of area, the overall average is 157 m ; FIS amount of trips (14%) and transport a small amount of establishments are, on average, 162.4 m ; while service- tonnage (5.5%). Tables 5, 6,and 7 present the descrip- related establishments are 158.2 m . The largest area per tive statistics of the different economic sectors for se- establishment is found in the education industry (288. 2 2 lected business attributes of Freight Intensive Sectors 4m ), followed by medium manufacturing (258.5 m ) (FIS)—industry sectors that generate more freight—, and transportation and warehousing (211.8 m ); while, Service Intensive Sectors (SIS)— industry sectors that the smallest area per establishment is found for mining 2 2 generate less freight and more services—,andallcom- (87.3 m ), fuel/lubricants distribution (85.9 m ), machine 2 2 mercial establishments, respectively. renting (83.9 m ), and utilities (53.3 m ). Although there As shown in Tables 5, 6 and 7, the economic sectors can is a large variance of business size within a single sector, be broadly classified as FIS (freight-related) and SIS these numbers shed light on the type of establishments (service oriented). The former includes establishments included in those sectors. where the main economic activity depends on cargo As expected, FIS establishments generate, on average, handling (e.g., manufacturing, wholesale, and retail more freight (freight attraction plus production) than Table 4 Breakdown of trips and tonnage by vehicle class Transportation mode Daily trips Daily Kilograms Average Kgs/trip Percent of trips Percent of tones All trucks 3882 267,546 68.9 64.4% 72.7% Pick-ups and vans 692 41,540 60.1 11.5% 11.3% Cars 491 35,490 72.2 8.1% 9.6% Motorcycles 626 17,162 27.4 10.4% 4.7% Bicycles 201 2811 14.0 3.3% 0.8% Others 138 3346 24.2 2.3% 0.9% Total 6031 367,896 44.5 100.0% 100.0% Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 8 of 11 Table 5 Statistics of Freight Intensive Sector (FIS) commercial establishments CIIU N Statistics Employees Est. area Freight attraction Freight production FTA deliveries FTP shipments Vendors (full and part time) (sq.m) (Kg/day) (Kg/day) (trips/day) (trips/day) Mining (CIIU 10–14) 39 Mean 4.23 87.26 51.27 4.44 0.91 0.19 5.46 Std.Dev 5.28 110.29 93.90 16.00 0.96 0.55 12.62 Min-Max 1–33 10–600 0.1–444.2 0–86.5 0–4.2 0–2.5 1–80 Light manufacturing 352 Mean 6.18 130.01 116.38 24.03 0.81 0.83 4.09 (CIIU 15–19) Std.Dev 14.40 276.81 710.10 148.75 1.27 5.63 7.02 Min-Max 0.9–220 1–2800 0.1–11,779.3 0–2538.5 0–13.8 0–100.7 1–70 Medium manufacturing 408 Mean 10.54 258.52 131.50 57.73 1.25 1.83 4.57 (CIIU 20–29) Std.Dev 31.32 1587.77 535.39 445.60 7.20 19.15 8.42 Min-Max 1–417 2–30,000 0–9038.5 0–8538.5 0–142.3 0–384.6 1–86 Heavy manufacturing 195 Mean 8.02 102.98 42.85 13.99 0.67 0.25 3.11 (CIIU 30–36) Std.Dev 19.38 156.10 176.12 92.74 1.22 0.77 4.28 Min-Max 1–150 3–1200 0.1–1765.4 0–1153.8 0–9.2 0–7.7 1–48 Recycle stores 22 Mean 9.92 188.50 179.78 339.91 0.70 0.42 2.32 (CIIU 37) Std.Dev 16.78 207.27 490.27 641.22 1.09 0.50 1.84 Min-Max 0.9–77 12–700 0–2269.2 0–2538.5 0–4.6 0–1.9 1–7 Construction (CIIU 45) 126 Mean 10.32 144.33 57.80 23.56 1.33 1.04 3.66 Std.Dev 20.42 245.39 112.17 94.31 7.63 4.01 6.04 Min-Max 1–200 4–1900 0.1–615.4 0–923.1 0–85.6 0–26.9 1–60 Repair stores (CIIU 50) 194 Mean 7.32 187.25 111.77 45.60 0.67 0.43 3.63 Std.Dev 14.94 607.59 389.97 203.04 1.00 1.54 5.68 Min-Max 1–137 2–8000 0–4043.5 0–2000 0–6.2 0–15.4 1–50 Fuel/lubricants 12 Mean 3.56 85.92 491.00 0.29 0.58 0.08 2.67 distribution (CIIU 505) Std.Dev 2.51 97.87 1169.64 1.00 0.44 0.27 1.72 Min-Max 1–98–300 0.5–4043.5 0–3.5 0.1–1.2 0–0.9 1–6 Wholesale (CIIU 51) 212 Mean 5.90 166.86 137.37 98.39 0.69 0.72 2.87 Std.Dev 8.93 434.60 511.13 674.34 1.13 3.52 2.94 Min-Max 1–80 6–4000 0–4769.2 0–8384.6 0–7.7 0–46.2 1–30 Retail (CIIU 52) 477 Mean 4.73 128.08 81.44 19.13 1.78 2.64 3.87 Std.Dev 10.07 690.58 501.27 135.83 21.13 42.73 8.54 Min-Max 0.9–80 1–4500 0.02–1765.39 0.04–253.85 0.04–9.23 0.04–26.92 1–70 Accommodation and 224 Mean 4.64 147.93 92.65 3.59 0.95 0.26 3.58 food (CIIU 55) Std.Dev 6.70 437.86 927.50 37.19 0.88 1.39 3.78 Min-Max 0.9–60 4–6000 0–13,882.3 0–542.3 0–6.2 0–13.5 1–45 Transport /warehouse 143 Mean 39.70 211.80 44.91 38.02 0.64 0.80 2.52 (CIIU 60–64) Std.Dev 350.96 848.24 184.99 217.91 1.16 2.40 4.01 Min-Max 1–4200 3–9800 0–1600 0–2384.6 0–7.7 0–19.2 1–30 Machine renting 49 Mean 3.00 83.90 16.02 24.41 0.63 0.22 2.52 (CIIU 71) Std.Dev 2.45 124.59 55.09 124.11 0.84 0.63 1.93 Min-Max 1–14 8–620 0–307.7 0–692.3 0–3.8 0–3.1 1–10 Notes: (1) The column labeled N corresponds to the sample size (number of observations); (2) Shipments excludes zero values; (3) Employment is calculated as full-time employees + 0.45 part-time employees; (4) FIS establishments generate on average 135 kg/establishment/day SIS ones, with 135 and 42.7 daily kg/establishment re- medium manufacturing have particularly high freight spectively. Among freight-related establishments, recyc- generation with 519.7, 491.3, 235.8, and 189.2 daily kg/ ling establishments, fuel distribution, wholesale and establishment, respectively. It is noteworthy that Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 9 of 11 Table 6 Statistics of non-freight intensive sector (service-oriented) commercial establishments CIIU N Statistics Employees Est. area Freight attraction Freight production FTA deliveries FTP shipments Vendors (full and part time) (sq.m) (Kg/day) (Kg/day) (trips/day) (trips/day) Utilities(CIIU 40–41) 3 Mean 4.48 53.33 2.17 0.04 0.13 0.05 2.00 Std.Dev 1.78 31.75 2.65 0.07 0.04 0.09 1.00 Min-Max 3–6.5 35–90 0.2–5.2 0–0.1 0.1–0.2 0–0.2 1–3 Financial services 67 Mean 17.01 197.94 12.38 4.06 1.27 0.40 3.30 (CIIU 65–67) Std.Dev 47.64 854.02 38.14 22.87 4.66 1.02 6.22 Min-Max 1–300 8–7000 0–278.3 0–184.6 0–30.8 0–4.6 1–40 Real state 210 Mean 8.84 90.32 17.32 15.23 0.58 0.29 2.58 (CIIU 72–74) Std.Dev 32.34 206.36 51.56 100.86 1.00 0.80 3.24 Min-Max 0.9–400 1–2500 0–338.5 0–1000 0–6.6 0–7.7 1–30 Education (CIIU 80) 47 Mean 17.13 288.45 14.85 9.83 0.49 0.89 3.23 Std.Dev 56.30 309.20 29.09 67.32 1.22 6.06 8.58 Min-Max 1–390 9–1600 0–126.9 0–461.5 0–6.9 0–41.5 1–60 Health and social 47 Mean 17.13 288.45 14.85 9.83 0.49 0.89 3.23 services (CIIU 85) Std.Dev 56.30 309.20 29.09 67.32 1.22 6.06 8.58 Min-Max 1–390 9–1600 0–126.9 0–461.5 0–6.9 0–41.5 1–60 Other services 54 Mean 12.05 152.04 159.02 4.41 0.38 0.08 3.00 (CIIU 90–93) Std.Dev 41.67 283.59 941.86 24.28 0.47 0.33 2.79 Min-Max 0.9–300 10–1600 0–6925.7 0–176.9 0–20–2.3 1–20 Notes: (1) The column labeled N corresponds to the sample size (number of observations); (2) Shipments excludes zero values; (3) Employment is calculated as full-time employees + 0.45 part-time employees; (4) Non-FIS establishments generate on average 42.7 kg/establishment/day establishments in mining, heavy manufacturing and numbers confirm that freight generation tends to be transport/warehousing are among the ones with the higher for freight-related establishments than for lower freight generation for freight-intensive sectors with service-related ones. 55.7, 56.8, 82.9 daily kg/establishment respectively. The In the case of freight trip generation (attraction plus latter could be a result of the type of activities covered production), FIS establishments generate, on average, by these sectors in the area of study. Mining covers ex- more freight trips than service-oriented ones, with 2.3 traction of stones and oil, and other mining support ac- and 1.0 daily trips/establishment respectively (more than tivities, which are not very common in Medellin. Heavy 2 times larger). Furthermore, the data shows that 70% of manufacturing covers motor vehicles, trailers and semi- establishments only receive shipments while about a trailers, machinery and equipment, and the like that is 30% also send shipments to their customers. Among the not very common in Medellin either, that could explain freight-related sectors, retail, medium manufacturing the low values of freight generation in those sectors in and construction have the highest freight trip generation the city. Those values are much higher in other places of with 4.4, 3.1, and 2.4 daily trips/establishment, respect- the country (e.g., el Cerrejón) and Medellin just receives ively; while heavy manufacturing, machine renting and the cargo (freight attraction) more than sending the fuel/lubricants distribution have the lowest with less cargo (freight production). than 1 daily trip. For service-related sectors, financial For SIS, average freight generation does not exceed 40 services, education and health and social services have daily kg/establishments with the exception of other ser- the highest freight trip generation with 1.7, 1.4 and 1.4 vices (163.4 daily kg/establishment). In essence, these daily trips respectively. It is noteworthy that some Table 7 Statistics of all commercial establishments Statistics Employees Est. area Freight attraction Freight production FTA FTP Vendors (full and part time) (sq.m) (Kg/day) (Kg/day) (trips/day) (trips/day) Mean 9.14 157.37 90.58 34.26 0.99 1.05 3.58 Std.Dev 80.45 743.24 520.32 283.49 9.12 18.79 6.39 Min-Max 0.9–4200 1–30,000 0–13,882.3 0–8538.5 0–461.5 0–923.1 1–170 Notes: (1) Shipments excludes zero values; (2) Employment is calculated as full-time employees + 0.45 part-time employees; (3) Sample size: 2947 establishments Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 10 of 11 sectors in the service-related sectors have higher freight complementary studies (e.g., large traffic generators, ser- trip generation than for freight-related sectors. This vice trips), and 3) Propose freight policies to improve could be to the use of small trucks (generating more freight transportation system in the Medellin metropolitan truck-trips) and not the amount of cargo transported. area. The project is currently underway and it is a proof This survey inquired about the number of vendors that starting to collect freight data and analyzing it could (small suppliers located in the CBD), which is not a lead the public sector to increase the interest in freight common practice for freight surveys. This information planning that obviously will help the development of can provide valuable insight on the type of activities urban areas, where freight is and will be a constant. taking place for each sector and to explain freight trip generation. As shown in the tables, the sampled estab- 6 Conclusions lishments have 3.6 vendors per establishment on aver- This paper seeks to analyze freight generation and freight age. Freight-related sectors have on average 3.7 vendors trip generation patterns in the metropolitan area of per establishment; while service-related sectors have 2.9 Medellin, Colombia; and to suggest how these freight pat- vendors per establishment. The sectors with more ven- terns should be considered for future planning decisions. dors are mining (5.5 vendors/establishment), medium In doing so, this study presents the basic freight data col- manufacturing (4.6 vendors/establishment), light manu- lection techniques used to depict all the complexity of facturing (4.1 vendors/establishment) and retail (3.87 freight distribution in metropolitan areas. The studies vendors/establishment). For service-related sectors, the consisted in an external cordon survey to catch the cargo averages are more homogenous with all of them having entering and exiting from the area; an establishment sur- between 2 and 3 vendors per establishment. vey applied to small and medium size commercial estab- These descriptive statistics highlight the differences lishment to study the internal movement of cargo; and a across sectors and depict the different freight and busi- series of interviews to some companies of carries and ness size patterns. One important observation is that trucks drivers with the objective of precise some logistics sectors with higher business size are not necessarily the practices and to have a more deep knowledge of the urban ones generating more freight or freight trips; also, sec- freight generation. All these surveys complement each tors with the highest freight generation are not necessar- other, and must be conducted to have a good depiction of ily the ones with the highest freight trip generation. This the freight movements in urban areas. can be explained by the different shipment sizes and the As expected, it was found that in the Medellin Metro- types of freight vehicles used in different sectors. politan Area, freight intensive sector establishments gener- ates, on average, significantly more cargo (freight 5.3 Planning implications attraction plus production) than the non-freight intensive Considering that 1) there is a profound lack of informa- sector (service-oriented) ones. Moreover, the authors tion about the amount of freight, freight trips and ser- found that the number of vendors (suppliers) of the com- vice trips in urban areas; and 2) this lack of information mercial establishments can be obtained thru freight sur- prevents that planners and transportation engineers veys (which is not a common practice). This information realize about the magnitude of the problem and find ad- can provide valuable insight on the type of activities taking equate solutions to freight issues; this study provides a place for each sector and to explain freight trip generation. good start of how can we shed light on decision makers The analyses of the surveys allow the decision makers to continue planning the cities, but now considering the to understand the nature of the cargo and the generation movement of goods in a more reliable fashion. Thus, patterns in different type of establishments. This this paper contributes to this gap in the in the quantifi- characterization of the freight patterns is vital for the cation of the problem based on a freight study and its forecasting of the behavior of the cargo and it is the findings. As an example of this, and based on the freight main input to perform freight demand modeling for city data collected in this study and its results, the local gov- planning, especially for developing countries, where ernment decided to start using the freight information there are too many budget constraints. to analyze projects, programs and policies to improve freight transportation system in the Medellin metropol- itan area. Although it took almost 5 years after this first study ended, in 2017 the Medellin Department of Trans- 7 Endnotes portation and the Area Metropolitana del Valle de A compendium of these strategies can be found in . Aburra (leader public institution that is the emission/ For more information about freight studies, please see  transportation authority in the area) decided to expand The data about trash and debris were obtained from this study in three scopes: 1) Update the freight OD information provided by the Metropolitan Area of the matrix; 2) Characterize the freight movements with Aburra Valley agency  Gonzalez-Calderon et al. European Transport Research Review (2018) 10:23 Page 11 of 11 For instance, the amount of freight moved per person 9. Holguín-Veras J, Thorson E (2003) Practical implications of modeling commercial vehicle empty trips. Transp Res Rec 1833:87–94 per day in New York City is about twice (50 kg/per- 10. Holguín-Veras J (2013) Freight demand modeling: state of the art and son/day) than for Medellin Metropolitan Area . practice, Presented at the adapting freight models and traditional freight It is important to note that the amount of freight data programs for performance measurement Workshop, Washington, D.C. Available: http://onlinepubs.trb.org/onlinepubs/conferences/2013/Freight/ moved per person per day (25 kg in Medellin) is dif- Holguin-Veras.pdf. Accessed 10 Nov 2016 ferent to the amount of freight consumed by a person per day (about 1–2 kg) that could be similar in differ- ent countries. Acknowledgements The research is based on the freight surveys that were funded by the Metropolitan Area of Aburra Valley and the Municipality of Medellin. The survey was conducted by the Universidad Nacional de Colombia at Medellin and advised by the VREF Center of Excellence for Sustainable Urban Freight Systems at the Rensselaer Polytechnic Institute. Authors’ contribution All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Author details Department of Civil Engineering, Universidad Nacional de Colombia at Medellin, Calle 65 # 78-28, M1-110 Medellin, Colombia. Department of Technology Management and Economics, Chalmers University of Technology, Vera Sandbergs Allé 8, Room 3332, SE-412 96 Göteborg, Sweden. Department of Civil Engineering, Universidad Nacional de Colombia at Medellin, Calle 65 No. 78-28, M1-223 Medellin, Colombia. Department of Civil and Environmental Engineering, Rensselaer Polytechnic Institute, 110 8th St. JEC 4030, Troy, NY 12180, USA. Received: 10 January 2017 Accepted: 5 April 2018 References 1. Holguín-Veras J, Jaller M (2014) Comprehensive freight demand data collection framework for large urban areas. In: González-Feliu J, Semet F, Routhier JL (eds) Sustainable urban logistics: concepts, methods and information systems. Springer-Verlag, Berlin 2. Meyburg A et al (2003) Special report 276: a concept for a national freight data program. Transportation Research Board, Washington, D.C 0309085705, Available: http://www.nap.edu/catalog.php?record_id=10793 3. Allen J, Browne M, Cherrett T (2012) Survey techniques in urban freight transport studies. Transp Rev 32(3):287–311 4. Área Metropolitana del Valle de Aburrá, Municipio de Medellín, and Universidad Nacional de Colombia Sede Medellín (2012) Encuesta Origen Destino de Hogares y de Carga para el Valle de Aburrá. Área Metropolitana del Valle de Aburrá, Medellín Available: http://www.metropol.gov.co/ Movilidad/Documents/Encuesta%20origen%20y%20destino%20de%20 hogares%202012.pdf 5. Holguín-Veras J, Ban J, Jaller M, Destro L, Marquis R (2010) Feasibility study for freight data collection - final report. Rensselaer Polytechnic Institute, Albany Available: http://www.utrc2.org/sites/default/files/pubs/Feasibility- Study-Freight-Data-Collection-Final.pdf 6. Holguín-Veras J, Sánchez I, González-Calderón C, Sarmiento I, Thorson E (2011) Time-dependent effects on parameters of freight demand models. Transp Res Rec 2224(1):42–50 7. González-Calderón CA, Sánchez-Díaz I, Holguín-Veras J (2011) An empirical investigation on the impacts of spatial and temporal aggregation of empty trips models. In: 4th National Urban Freight Conference. METRANS, Long Beach, p 17 8. Holguín-Veras J, Thorson E (2003) Modeling commercial vehicle empty trips with a first order trip chain model. Transp Res B 37(2):129–148
European Transport Research Review – Springer Journals
Published: Jun 5, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera