Extract Most Frequented Locations from individual spatio-temporal trajectories

Description

This script returns the location most frequented by an individual (called MFL) during weekdays' daytime and nighttime according to a certain time window. The MFL during a given time windows is defined as the location in which the individual has spent most of his/her time. In this algorithm, two scales are considered, hours and days. Hopefully, for a given individual, the MFLs detected at the two scales should be the same.

Input

The algorithm takes as input a 6 columns csv file with column names, the value separator is a semicolon ";". Each row of the file represents a spatio-temporal position of an individual's trajectory. The time is given by the columns 2 to 5 (year, month, day and hour). It is important to note that the table should be SORTED by individual ID and time.

1. ID of the individual
2. Year
3. Month (1->12)
4. Day (1->31)
5. Hour (0->23)
6. ID of the geographical location

Parameters

The algorithm has 6 parameters:

1. wdinput: Path of the input file (ex: "input.csv")
2. wdoutput: Path of the output file (ex: "outputoftheawsomemaximelenormandsalgorithm.csv")
3. minH: Lower bound (included) of the night time window (ex: 20h)
4. maxH: Upper bound (included) of the night time window (ex: 7h)
5. minW: Lower bound (included) of the day time window (ex: 8h)
6. maxW: Upper bound (included) of the day time window (ex: 19h)

Output

The algorithm returns a 19 columns csv file with column names, the value separator is a semicolon ";". Each row represents an individual:

1. ID: ID of the individual
2. NbMonths: Number of distinct months covered by the trajectory
3. NbConsMonths: Maximum number of consecutive months covered by the trajectory
4. MFLHomeDays: MFL during nighttime (day scale), 'NoMFL' if NbDaysHome=0
5. MFLHomeDays2: Second MFL during nighttime (day scale) if ex aqueo, 'NoMFL' otherwise
6. NbDaysHomeMFL: Number of distinct days (during nighttime) spent in the MFL
7. NbDaysHome: Number of distinct days covered by the trajectory (during nighttime)
8. MFLHomeHours: MFL during nighttime (hour scale), 'NoMFL' if NbHoursHome=0
9. MFLHomeHours2: Second MFL during nighttime (hour scale) if ex aqueo, 'NoMFL' otherwise
10. NbHoursHomeMFL: Number of distinct hours spent in the MFL (during nighttime)
11. NbHoursHome: Number of distinct hours covered by the trajectory (during nighttime)
12. MFLWorkDays: MFL during daytime (day scale), 'NoMFL' if NbDaysWork=0
13. MFLWorkDays2: Second MFL during daytime (day scale) if ex aqueo, 'NoMFL' otherwise
14. NbDaysWorkMFL: Number of distinct days spent in the MFL (during daytime)
15. NbDaysWork: Number of distinct days covered by the trajectory (during daytime)
16. MFLWorkHours: MFL during daytime (hour scale), 'NoMFL' if NbHoursWork=0
17. MFLWorkHours2: Second MFL during daytime (hour scale) if ex aqueo, 'NoMFL' otherwise
18. NbHoursWorkMFL: Number of distinct days spent in the MFL (during daytime)
19. NbHoursWork: Number of distinct hours covered by the trajectory (during daytime)

Execution

You can run the code using the command:

python MFL.py input.csv output.csv 20 7 8 19

Citation

If you use this code, please cite:

Lenormand M, Louail T, Barthelemy M & Ramasco JJ (2016) Is spatial information in ICT data reliable? In proceedings of the 2016 Spatial Accuracy Conference, 9-17, Montpellier, France.

If you need help, find a bug, want to give me advice or feedback, please contact me! You can reach me at maxime.lenormand[at]irstea.fr