Trip Generation Estimates for Spatial Units based on Built Environment and Geocoded Household Travel Surveys

• Applicant: Professor Dr.-Ing. Tobias Kuhnimhof
• Subject Area: Traffic and Transport Systems, Intelligent and Automated Traffic ; Urbanism, Spatial Planning, Transportation and Infrastructure Planning, Landscape Planning
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 524931623

TRIGGSY aims at a novel approach to estimate trip rates, specifically the number of terminating trips per spatial unit. Such trip rates form a key input for numerous research and planning uses. However, existing trip generation estimation approaches either require laborious data collection (e.g. FGSV or Trip Generation Handbook approach) or an elaborate modelling process (e.g. travel demand model). The TRIGGSY-approach builds on a) geodata which is available open source in Germany as in many other countries and b) geocoded household travel survey data as is available from the latest German national travel survey “Mobilität in Deutschland 2017” (MiD). The TRIGGSY-approach can be applied nationwide for Germany to estimate trip rates and transferred to other locales with similar data. In the MiD (~1 million recorded trips by ~320,000 survey participants representing travel on an average day) there are 20.000 geographic grid cells (1 km by 1 km) for which trips terminating there have been recorded (return home trips excluded). For a given grid cell x, the number trips recorded in the MiD is influenced by two factors: I. Built environment and socio-economic factors of x and its environment; these determine the number of trips ending in x in reality (i.e. the universe of trips ending in x per day). II. Characteristics of the MiD-sample; naturally, the number of trips recorded ending in x (i.e. the respective sample of trips ending in x) is subject to the number of respondents interviewed in the vicinity of x. TRIGGSY aims at disentangling these two factors in order to establish a statistical relationship between the cell characteristics as the explanatory variables (factors I) and the rates of terminating trips as the explained variable. Poisson-regression is a possible approach, with the number of terminating trips recorded in the MiD scaling linearly with the survey respondents in the vicinity of the respective cell (i.e. the accessible survey respondents). Another possible approach is linear regression, e.g. when using the log of the recorded trips per respondent living in the vicinity (i.e. the trips per accessible respondent). Hence, the number of persons with accessibility to a respective cell in a) the MiD sample and b) the population plays a major role. Therefore, TRIGGSY will evaluate different measures of accessibility. Thus, TRIGGSY develops a model to estimate trip rates per accessible person for small geographic units based on their land use and additional spatial properties, e.g. trips generated per medical practice or per square meter of retail. Combining this with geodata on the accessible residential population allows for nationwide extrapolation of total trip rates. The project validates the TRIGGSY-approach through comparison with existing methods for estimating trip generation and publishes estimated trip rates online.

• DFG Programme: Research Grants