Repository for everything we use to win SOČ in 2017.<!--- Our work is focusing on sentiment analysis of public opinion on big world problems, like terrorism and migration. We are using social networks (Twitter) and well known discussion forums (Reddit, 4chan).--->
All code is written in python3.5.2.
We are using python library tweepy for accessing Twitter API.
We are using python library praw for accessing Reddit API.
We are using python library basc_py4chan for accessing 4chan API.
For initial analysis are using out-of-the-box python library textblob for both work with text (tokenizing, word tagging, ...) and sentiment analysis. For deeper analysis we are using custom sentiment classifier written in python with library nltk. We have two classifiers, each has same architecture, but is trained on different dataset. One is trained on labeled movie reviews. Second is trained on labeled tweets dataset. The second one is preferable for our purpose.