Fig. 1. Hourly variation of origin and destination points extracted from smart card transactions and taxi GPS trajectories

Fig. 2. The correlation between origin and destination time series oftwo traffic flowson weekdays and weekends

Fig. 3. Spatial distributions of average daily trips extracted from smart card transactions and taxi GPS trajectories

Fig. 4. The hot and cold spots of average daily trips extracted from smart card transactions and taxi GPS trajectories

Fig. 5. Histograms of log odds ratio based on each combination of trip category (outgoing vs. incoming) and type of day (weekdays vs. weekends)

Fig. 6. Spatial patterns of log odds ratio based on each combination of trip category (outgoing vs. incoming) and type of day (weekdays vs. weekends)

Fig. 7. Spatial distributions of the regression coefficients calculated for three independent variables

Fig. 8. Average travel distances of the trips extracted from smart card transactions and taxi GPS trajectories

Fig. 9. Histogram of the average travel distance extracted from the two traffic flows

Fig. 10. Distance decay of trips extracted from smart card transactions and taxi GPS trajectories on weekdays and weekends

Fig. 11. Spatial communities discovered from smart card and taxi data on weekdays and weekends