Here we developed the model in two steps. First we crawled twitter data based on manual keyword. Then we used clustering method to detect opioid related term by clustering method. We did the TFIDF to find the TFIDF values. Then we ranked the twitter data and choose the top ranked twitter. In second part we deployed the model on web server.
We didn't include all the raw data as there are lot of files. Only three files included as samples.