Querying Term Associations and their Temporal Evolution in Social Data

There is a growing number of applications that rely on data collected from online social networks. These applications typically issue search requests for keywords and process the data returned by online social networks through APIs. The selection of keywords can have an important impact on the quality of the results and the appropriateness of the collected data for further analysis. Indeed, adding or removing keywords in the search requests may change the characteristics of the sampled data. Hence, it is important for users to have the ability to explore data and to express complex requests in order to discover the context of collected data. In this work, we propose a model and a number of query operators that allow users to select data and explore its context by means of querying for associations between keywords or entities as well as their evolution over time. The model supports different time granularities and the calculation of term association weights based on the context of terms. We demonstrate the use of the model and the query operators with a running example based on data we have collected from the microblogging service Twitter, and a first implementation running on top of a relational database.