|M.Sc Student||Pat Barak|
|Subject||Geosocial Search: Finding Places based on Geotagged|
|Department||Department of Computer Science||Supervisors||Professor Joseph Gil|
|Dr. Yaron Kanza|
Geographic search - where the user provides keywords and receives relevant locations depicted on a map - is a popular web application. Typically, such a search is based on static geographic data. However, the abundance of geotagged posts in microblogs such as Twitter and in online social networks like Instagram, provide contemporary information that can be used to support geosocial searches - geographic searches based on user activities in social media. Such searches can point out where people talk (or tweet) about different topics. For example, search results may show where people refer to “jogging” to indicate popular jogging places. The difficulty of implementing such a search is that there is no natural partition of the space into “documents” as in ordinary web searches. Thus, it is not clear how to present results and how to rank and filter results effectively. In this thesis, we demonstrate a two-step process of first, quickly finding relevant areas by using an arbitrarily indexed partition of the space, and second, applying clustering on discovered areas to present more accurate results.
We propose four different ranking measures for evaluating the relevance of an area for a given query. The computation of those measures are supported by a hybrid partition-aware inverted index that we developed, to increase the efficiency of online searches. Additionally, we introduce a framework that utilizes geotagged posts in geographic searches and we illustrate how different ranking methods can be used based on the proposed two-step search process.
We validated the effectiveness and usefulness of geosocial search in three different ways. First, we measured the ability to find events in the NYC area. Second, we compared the query processing time of the different proposed algorithms for computing the different semantics.Third, we conducted a qualitative analysis of the effectiveness of geosocial search against ordinary geographic search.
Our experiments show that geosocial search is more effective than ordinary geographical search when searching for locations of events, locations of activities, and locations that refer to an abstract sentiment, especially when those locations are not associated with a predefined geographical entity.