This mini-project presents the workflow of delivering a Machine Learning model responsible for Twitter sentiment analysis. The goal is to classify posts into negative or positive. Some Exploratory Data Analysis together with model training was shown. For interacting with the model, a Streamlit app was created, which enables the user to input the Twitter Post, and the model will predict a sentiment for it.
The data used in this mini-project was obtained from Kaggle, here. The dataset consists of labelled tweets, 800,000 posts for each class (negative/positive).
A simple LSTM model was implemented for sentiment analysis. Tensorflow library was used for implementing the model.
To launch a Streamlit app, run:
streamlit run streamlit.py
You may check some of my other projects here.