Title:  Building Adaptive Location-based Services using the NYC TLC Dataset

Description:

Mobile services which provide location based information on request of the user have in the  last few years become established parts in the daily life of many people. Services which, on
the contrary, provide information proactively, that is without a preceding user request, are an  exception. Although the technical hurdles for services such as these could be to the greatest
extent overcome, the difficulty still remains in deciding when a user should be provided with  which information. Through the analysis of large amounts of historical location data, insights
can be gained which could become available to a user in a specific context. On grounds of the  scope and heterogeneity of spatiotemporal data, the production and proactive supply of such
findings is a big challenge. Decisive therein is that often there is just a short time-frame in  which the information is relevant to the user. The present thesis suggests an approach which
should enable the link between location analytics and proactive location services. In order  to check the feasibility of the concept, a prototype for mobile usage was developed, with its
list of requirements being based on a questionnaire given to professional taxi drivers in Berlin.  Through the analysis of historical data from taxi drivers in Manhattan, New York City, an index
of attractiveness of individual areas could be deduced. For this purpose STEAM, a platform  for spatiotemporal analysis of heterogenous spatiotemporal datasets, was used. Based on a
user’s location, the gained information can proactively be made available. For this, the concept  of geofencing was implemented. With the help of the prototype, it could be shown that, with
some assistance, the optimization is viable. The reliability of the newly supplied information  was emphasized as problematic and thus has to be analyzed in a field trial.

Supervisor: Bersant Deva, Philip Raschke

Type:  Bachelor Thesis

Duration: 6 months