Development of a monitoring dashboard for sentiment and emotion in geolocated social media

Abstract

This thesis comes from a project which goal is to design and develop a dashboard with a graphic interface that shows information about the analysis of tweets posted in a geographic zone. The dashboard's main element is a heatmap that shows the value variation of certain sentiment parameters with a color gradient. Furthermore, this map will also clasify emotionally all analyzed data. Besides the map, the dashboard has to be ready to host more elements that may increase its functionality. This project offers two: a list of all tweets being displayed, and a component which allows filtering the dataset by date. All this elements require getting and preparing a large amount of data, which is also covered in this document. Since tweet fetching until dashboard data provision, a batch processing pipeline that makes several transformations to data has been de?ned. This is because it is interesting to simplify the web's workload and imporve user interaction speed. It is worth mentioning that this development has been made following the architecture defined by Sefarad project, mainained by Grupo de Sistemas Inteligentes. Finally, there are more topics included in this document, such as final conclusions, learnt technologies, problems found during it, and possible future lines of work.