Learning from our movements – Mobility data analytics

Yannis Theodoridis | Tuesday, July 17
Professor of Data Science at the Department of Informatics, University of Piraeus (Greece).


From raw location recordings to mobility patterns – how can we exploit on the ubiquitous GPS technology in order to get knowledge about our movement behavior? Which are the most representative examples of mobility patterns that can be mined from humans’ mobility datasets? In this course, issues and solutions on Mobility Data Analytics (MDA) are overviewed, including acquisition, storage, processing, and mining aspects. Current trends in MDA, such as semantic / holistic trajectory modelling and big mobility data management, are also presented. Theoretical modules will be interleaved with hands on parts. Use cases include urban, sea, and/or air space movement.

Course outline:

1.Getting to know your data
– Nature of data; sources; applications
– Measuring data (dis-)similarity

2.Pre-processing your data
– Tasks at individual level (Data cleansing, transformation, simplification, enrichment, etc,)
– Tasks at dataset level (Data storage, indexing, etc.)

3.Analyzing your data
– Discovering groups and collective behaviours (Cluster analysis)
– Discovering usual vs. unusual patterns (Frequent pattern vs. Outlier detection analysis)
– Forecasting anticipated movement (Classification and Prediction)

– An end-to-end use case
– What’s next: a near-future research agenda