This space hosts projects involving data scrapping, processing raw data, feature extraction, predictive modelling purely for learning/exploring purposes. These projects have been taken up in personal capacity and I have solely worked on these projects. Any sources/papers which have been referred have been accordingly citied in the project files.
The projects have been listed below
Movie Revenue Prediction - This project consisted of building a predictive model that could be used to estimate the box office potential of a movie, given its title, genres, runtime, plot & other features. The trained regression model was tested on an unseen dataset and the top 30 predictions of the model had a deviation less than 5 millions USD.
IPL Score Prediction - This project is a work in progress project. As a part of this project, a predictive model is being built which can predict the score of an innings (20 overs) given the score at the end of 5 overs of that innings. The IPL macthes are of 20 overs duration and highly unpredictable. Given the data that is available at our disposal - 10 teams, 10 years of data (approx. 1500 innings), this does make for an interesting use case for predictive modelling. This project does provide a lot of scope for data preprocessing/engineering which is a precusrsor to any data science projects in real world.
Sudoku Solver - This project consists of an executable file built using Python which can solve a traditional Sudoku puzzle. The end goal of this project is to have a Machine Learning model which could be capable of solving a sudoku puzzle. Having a code based solution provides a first step in this direction.
COVID 19 Forecast - This project is about forecasting COID 19 cases & deaths for 180 countries. This competition is hosted on kaggle on weekly basis. Here an attempt has been made to develop prediction model using the Week 4 COVID 19 dataset.