Paper Info
Title
Mobility data mining: discovering movement patterns from trajectory data
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.
Year
DOI
Venue
2010
10.1145/1899441.1899444
Computational Transportation Science
Keywords
Field
DocType
movement data,movement pattern,storing location-aware data,location data,semantically rich data,trajectory data,mobility data mining,spatio-temporal data,poor data quality,mobility data,raw data,high-level collective mobility knowledge,human mobility,data mining,data quality,expert system,mobile device,wireless network,data collection,real time
Data mining,Obsolescence,Data quality,Statistic,Mobility model,Raw data,Mobile device,Global Positioning System,Geography,Semantics
Conference
Citations 
PageRank 
References 
7
0.44
1
Authors
7
Name
Order
Citations
PageRank
Fosca Giannotti12948253.39
Mirco Nanni2141284.47
Dino Pedreschi33083244.47
Fabio Pinelli497250.96
Chiara Renso592576.04
Salvatore Rinzivillo667344.49
Roberto Trasarti771045.82