Description of the Data

The dataset to be released is called ‘People Flow Data’, which consists of spatio-temporal data of particles processed from Person Trip Survey Data (PT Data). The origin survey is collected by a local authority of each area, aims at studying the actual daily travel behavior in cities with a large population (1%~3% sampling rate). The dataset provides individuals’ minute-to-minute location information based on detailed network data. We open the whole dataset in the Tokyo area as a training dataset for participants to model and calibrate the developed model. For the test dataset, a synthetic population contain demographic attributes as age, gender, and occupation and their home location is provided.

Table 1. Description of the dataset
Area Population Survey Date Transport Mode Attribute
Training Tokyo About 600,000 2008/10/01 17 type age, gender, occupation, purpose
Test Osaka About 400,000 2010/10/01

Table 2. Code of transport mode
Code Content Code Content
1 Walk 7 Minicar
2 Bicycle 8 Van
3 Motorized two-wheel vehicle 9 Bus
4 Motor bicycle 12 Train
5 Taxi 97 Stay
6 Vehicle 99 Unknown

Table 3. Code of occupation
Code Occupation Code Occupation
1 agriculture 9 Engineer
2 mining 10 Manager
3 Craftsman 11 Others
4 Sales 12 Student (Under15)
5 Service 13 Student (above 15)
6 Transportation 14 Housewife
7 Security 15 No job
8 Office clerk 16 Unknown

More details of the dataset are shown in (

Conditions of use of the dataset

Considering privacy problems, PT data sets are processed carefully to avoid reproduction from our providing data sets. However, the dataset is not totally open. Any use must comply with the following:

  • Participants may download, use and copy the dataset only for this competition. Any personal use, commercial purpose and supply the material to any other person are forbidden.
  • Any procedure that using the provided dataset to identify specific person is prohibited.

The dataset is available for joint research apart from this competition. CSIS (Center for Spatial Information Science, University of Tokyo) deliberative council will check the research proposal (may take about a week), and then send you approval notice. For more details, please check (