At EndemolShine, employees were interested in finding the trends that are present among viewers of their TV productions and wanted to measure the feelings of viewers on certain topics.
We built connectors to periodically obtain data from several social media channels in a single data lake environment.
We have then applied an algorithm to filter irrelevant messages such as advertisements from this data. On all the remaining relevant messages, we applied a sentiment analysis algorithm. Through entity analysis we could map subjects and in this way we created insights on public opinion on specific topics.