Soda - Social Media Data Analysis

Most of today’s commercial companies rely heavily on social media and community management tools to interact with their clients and analyze their online behaviors. Nonetheless, these tools still lack evolved data mining and visualization features to tailor the analysis in order to support useful marketing decisions. We present an original methodology that aims at formalizing the marketing need of the company and develop a tool that can support it. The methodology is derived from the CRISP-DM approach and includes additional steps dedicated to the design and development of visualizations of mined data. It is based on two Swiss companies use cases. First, a prototype that aims at understanding the needs of tourists based on Instagram data. We extend the existing state-of-the-art by enriching hashtags analysis methods with a semantic network based on Linked Data. Second analyzes internal customer data of an online discount retailers to help them define guerilla marketing measures.