The codes can be accessed from here.
The python codes SSTM.py has been developed to generate the SSTM. Specifically, it derives the visit frequency of the track points based on each user’s media access control (MAC) address.
Inputs: Input data for running the SSTM.py include the raw Wi-Fi data (filename: Wifi.csv in the zipped file wifi.zip), weekend dates during the study period (filename: weekend_date_dict.xlsx), time periods in a day (filename: time_period_dict.xlsx), and the major activity type near each Wi-Fi detector (filename: activity_type_dict.xlsx).
Please change the reading and writing file paths to your local directory before running the codes. It may take a few hours to run the codes.
Outputs: The output of running the SSTM.py is a csv file (output_SSTM.csv in the zipped file output_SSTM.zip) including ID, user_MAC, and two SSTMs (weekday and weekend, 4×4 for each). In the output_SSTM.csv, the kmeans_cluster column is the kmeans++ clustering result of the SSTMs, and this column is derived by the MATLAB codes kmeans_SSTM.m.
Abbreviations (in the output_SSTM.csv):
Day type: W1: weekday, W2: weekend;
Time period: TP1: 0:00–9:00, TP2: 9:00-20:00, TP3: 20:00–22:00, TP4: 22:00–24:00;
Major activity type: AT1: tourist, AT2: residential, AT3: commercial, AT4: mixed;
kmeans_cluster: 1: Local resident, 2: Nearby resident, 3: Passing commuter, 4: Frequent visitor, 5: Weekday visitor, 6: Weekend visitor.