The Observatory

The Food Sentiment Observatory Tool (FoObs) is a multipurpose data exploration tool for social media data. The tool is designed to handle various social media inputs including online articles, social networking posts, blogs, etc. The current implementation provides the ability to collect and analyse collections of data from the Twitter social networking platform.

The initial FoObs prototype is under development. Below is a brief summary of currently implemented functionalities.

One of the main goals of the tool is to automatically record provenance of research activities broadly encompassing the three main steps of a social media research process namely data collection, data preparation and data analysis. Foobs will implement an extended version of the PRISM data model for recording provenance in an interchangeable machine-readable format using semantic web technologies.

Keyword search and frequency analyser provides a powerful and intuitive ability to explore and create interesting subsets of collected data. Such data collections can then be further analysed by other modules from the FoObs toolkit.

Clustering analysis can be applied to create groups of social media messages containing keywords of interest (e.g. mentioning a food item) and further analysed with additional list of terms (e.g. references to sickness).

A two dimensional scatterplot analysis for comparison of two dataset groups is enabled via integration of the ScatterText library. The library generates an interactive visualisation of frequent terms use within the two groups. In a addition, a list of most frequent terms and the ability to browse the raw messages is also provided by the tool.