Computer vision-enriched discrete choice models

This repo contains the data that are used to train the computer vision-enriched discrete choice models (CV-DCM) proposed in Van Cranenburgh & Garrido-Valenzuela (2024).

Data

The file "data_CV_DCM.csv" contains data from a stated choice experiment. In the experiment, respondents were presented two residential location alternatives, and were asked to indicate which alternative they would choose. Both alternatives comprise travel time (TT), monthly housing cost (C) and an image showing the street-level conditions. Here, you see a screenshot of a choice task from the stated choice experiment.

The table below lists the most important variables in the data set:

Variable	Description
CHOICE	1 if the respondent chose alternative 1, 2 if the respondent chose alternative 2
Ci	Monthly housing cost of alternative i
TTi	Travel time to alternative i
IMGi	Image ID used of alternative i
IMG_LATi	Latitude of the location of the image of alternative i
IMG_LNGi	Longitude of the location of the image of alternative i
ANGLE_FROM_NORTH_IMGi	Angle from the north in degrees used to take the image of alternative i
AGE	Age of the respondent
GENDER	Gender of the respondent
PROVINCE	Province of the respondent
POSTCODE	Postcode of the respondent
train	1 if the respondent was in the training set
test	1 if the respondent was in the test set

More details about the data collection can be found in section 3 of the associated paper: Van Cranenburgh & Garrido-Valenzuela (2024)

License CC BY-NC-SA 4.0

Shield:

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.