Hashtag Emotion Chart Visualization

Ever wanted to get an overall feel of how a trending hashtag is doing on Twitter? Or maybe track how people are reacting to the hashtag of your marketing campaign? Just visit {INSERT LIVE LINK} and enter the hashtag that you want to track the emotion of. It may take a 5-10 seconds to gather the tweets, but you will soon be presented with a pie chart visualization behind the emotions of what people are feeling behind that hashtag.

How it works

Using Tweepy, Twitter's data streaming API, in order to obtain the tweets that utilize the hashtag given by the user. All collected tweets are then saved to a CSV file and are sent to be preprocessed. Once preprocessed, the tweets are run through a neural network MLP classifier with a word embedding layer into 13 possible labels (emotions):

• Anger
• Boredom
• Empty
• Enthusiasm
• Fun
• Happiness
• Hate
• Love
• Neutral
• Relief
• Sadness
• Surprise
• Worry

Installing and Usage

To utilize this application, visit: https://hashtag-emotion-chart.herokuapp.com/ and enter the hashtag you wish to track the emotion behind.

For local usage run the following commands:

• git clone https://github.com/natepill/Hashtag-Sentiment-Tracker.git
• cd Hashtag-Sentiment-Tracker
• pip install -r requirements.txt
• python app.py

Next Steps

• Given multiple hashtags to track, not just a single one
• Add user accounts so users may save and reference back to prior tracking sessions
• Give users dashboard analytics of all the metadata from the people using the hashtag they entered provided by the Tweepy API.

Built With

• Flask
• Tweepy API
• ChartJS
• Python
• Keras
• Sklearn
• nltk

License

This project is licensed under the MIT License