The question behind this project was simple to ask but hard to answer with confidence: can you predict whether a book is likely to be adapted into a successful movie before it happens? Studios and publishers make these bets constantly, often on instinct. This project explored whether sentiment data and audience response could make that decision more data-driven.

Rather than relying on a single data source, the approach combined multiple datasets and sentiment analysis techniques to build a model that could actually generalize, not just fit one example well.

The approach included:

Pulling and cleaning data from 4-5 separate data files, plus scraping IMDB review data directly

Using TextBlob for sentiment analysis to quantify audience reaction to both books and their film adaptations

Testing two distinct modeling approaches to answer two different questions: one predictive and one classificatory

Preprocessed and merged 4-5 data files alongside scraped IMDB review data

Applied TextBlob sentiment analysis to quantify book and movie audience reception

Built a linear regression model to estimate a movie's likely rating based on its source book's rating

Built a logistic regression model to estimate the likelihood of a book successfully turning into a movie

Validated findings through feature importance analysis and sentiment comparison visualizations across the top 10 movies with the greatest book-to-film differences

This project shows the ability to go beyond standard marketing analytics into applied data science by building and validating an actual predictive model, not just reporting on existing data. It demonstrates the kind of technical depth that lets data-driven decisions go further than dashboards and reporting alone.

An 88% accuracy model that can help predict whether a book is likely to be adapted into a successful movie, using its book rating, number of ratings, and number of text reviews as key features.

Tools: Python, Google Colab, TextBlob, Linear Regression, Logistic Regression