• Scraped Democrats and Republican tweets using Twitter API, processed and cleaned tweets to build an animated political persona.
• Trained a bidirectional LSTM model using GloVe Embeddings that predicts the political ideology score of text gives interest scores on a scale of 0 to 1.
• Obtained 95% cross-validation accuracy and Performed lift analysis to show the important terms in the text to improve the interest score.
• Prepared visualizations, compared and contrasted the original model score and updated scores to suggest add/remove word in the given text.