Mobility data mining

Abstract

The analysis of movement data has been recently fostered by the widespread diffusion of new techniques and systems for monitoring, collecting and storing location-aware data, generated by a wealth of technological infrastructures, such as GPS positioning and wireless networks [2]. These have made available massive repositories of spatio-temporal data recording human mobile activities, such as location data from mobile phones, GPS tracks from mobile devices, etc.: is it possible to discover from these data useful and timely knowledge about human mobility? The GeoPKDD project [1], since 2005, investigated this direction of research; the lesson learned is that there is a long way to go from raw data of individual trajectories up to high-level collective mobility knowledge, capable of supporting the decisions of mobility and transportation managers. Such analysts reason about semantically rich concepts, such as systematic vs. occasional movement behavior and homework commuting patterns; accordingly, the mainstream analytical tools of transportation engineering, such as origin/destination matrices, are based on semantically rich data collected by means of field surveys and interviews. Clearly, the price to pay for this richness is hard: mass surveys are very expensive, so that their periodicity is very broad and obsolescence is rapid; poor data quality is also a plague: people tend to respond elusively and inaccurately. On the other extreme, automatically sensed mobility data record individual trajectories at mass level, in real time. Clearly, the price to pay here is exactly the lack of semantics in raw data: How to bridge this deficiency? Current generation of analysis systems and methods on mobility data are based on standard statistical techniques, such as movement distributions among the city areas. However, we claim that these are not sufficient to express movement oriented analysis. For example, detecting subgroups of users having similar movements is a behaviour not captured by standars statistic techniques, due to the strong variability among individuals movements.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 12 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.