Automatic summarization and visualisation of healthcare tweets

With social media encompassing people in all aspects of life, the relevance of information shared over these media is becoming highly relevant. The marketing and retail industry have been using social media like Twitter and Facebook extensively to collect information and promote their products. Now, it is the healthcare industry's turn to find hidden insights from the vast data available on the web. Analysing healthcare related tweets is one such arena which is proving to be useful in finding a lot of information about healthcare facilities like symptoms, diagnosis, treatment and recovery related to specific chronic diseases. As the data generated is huge, a summarized representation of the same is an important and non-trivial requirement which needs to be explored. In this work, we concentrate on analysing tweets related to cancer. We propose a simple clustering based method to summarize the Twitter generated data and extract more relevant tweets. In addition to this, two feature selection methods have been explored and the results compared which reiterates the importance of proper feature selection techniques in the process of Data analysis. The automatically generated summaries are evaluated using Cosine similarity measure. We also visualize the most frequent words using a Word cloud visualisation technique. Our work gives positive results and emphasises the fact that summarization of tweets provides a better understanding of the underlying data which can benefit healthcare industry which is aiming to provide personalised medicine for individuals.