Social media has been one of many reasons why young people have suffered from depression, anxiety, and self-confidence. Out of all the platforms, Twitter by far one of the biggest active media channels since it's texted based content stream makes it cheap and easy for tweets to be posted regarding the contexts. Because of the uncontrollable wide range of tweets, there would be hard for someone to notice negative tweets from other twitter's users even from closed ones. That is why we came up with the idea of detecting negative tweets, keep track of how many time a particular user would post depression tweets and display their current emotional state. We built it with MySQL for database integrated with Microsoft Azure database service, Python, Microsoft Azure using AI service for analyzing the sentiment of the tweets.
One of the hardest challenge that we ran into was the Twitter OAuth, difficulty in team participation, problems with connecting MySQL with Microsoft Azure with Python, integrating Flask with MySQL and Python.
We were able to get text from Tweets of hard-coded users and send text to Microsoft Azure Text-Analysis sentiment API to get sentiment score. We then added those entries into a Microsoft MySQL database.
What I learned
- How to both design, create, and query from a database
- Work w/ AI API (specifically Microsoft Azure)
- Work w/ APIs in general
What's next for Sentalyze
- Send texts through Twilio API (notify user if sentiment score too low)
- Graph sentiment scores over time (in real-time)