Roadmap

This document explains the functionality planned for various releases.

v2.0

✔️ Parallelization.
More method to predict activity without context data. #213
Add context data to movement trajectories (using spatio-temporal map algebra). #63
✔️ Implement tours (starting and ending at a persons home location). #287
✔️ Speed calculation for triplegs/positionfixes. #191
CodePeerReview: Next milestone after 1.0. #86
Include the calculation of common mobility indicators (e.g., radius of gyration)

I/O:

Preprocessing:

Short walks to/from cars/buses/etc.: These are often not recognized by the tracking applications.
Imputation of trivial gaps in tracking data (triplegs and staypoints).
Simple outlier filtering for triplegs
Create a solution to match trajectories from different tracking sources. #49

Augment tracking data:

Better transport mode prediction based on movement data (speed, associated features from accelerometer data, etc.) as well as by aligning them with context data such as from OpenStreetMap or GTFS departure schedules.
Better activity prediction. This functionality is based on properties of the movement data and on context data such as points of interests.
Provide holistic inference of transport modes (take into account the fact that if someone uses the car to reach a certain location, the person is likely to leave by car as well).
Map match triplegs (based on transport mode identification). Solutions could be LeuvenMapMatching or based on osrm.

Analysis

Mobility behaviour (and its changes): automatic detection of uncharacteristic changes in mobility patterns.
User profiling and clustering.
Anomaly detection (based on properties of movement data as well as contextual factors).
Clustering of triplegs using similarity metrics
Next place prediction (e.g., markov-model)

Visualization

Visualize tours geographically.
Visualize trips and locations geographically, e.g., by coloring triplegs differently or by making a buffer around all staypoints that are part of a locations.

Various