Social media is a blessing if we want to reach a larger audience but for people belonging to marginalised and oppressed part of society, it can be a frightening experience. Hate speech and humiliating someone on platforms is quite prevalent these days. We have made this application to detect your tone of text before you post it on social media so that it does not affect any person in a wrong way.

What it does

Text Tone Detector is a machine learning based application that detects the tone of the message as positive or negative to prevent hate speech.

How we built it

We built it using Jupyter Notebooks, NLTK,scikit-learn libraries , created the interface using streamlit and deployed using heroku.

Challenges we ran into

The major challenge was to improve accuracy of prediction. For this we preprocessed the data and refined it to reduce the amount of useless words , phrases and symbols. We also used porter stemmer to make changes in the words and convert them to their root form.

Accomplishments that we're proud of

We are proud of being able to put a fully functional application at the end of this hackathon. This project gave us an opportunity to refine our skills in ML.

What we learned

We learned NLP and data preprocessing. We also learned the use of regular expressions in removing unwanted parts of a string.

What's next for Text Tone Detector

  1. Better accuracy
  2. Improved UI
  3. Assistance for specially abled people
Share this project: