Data Science Projects

This repository contains a collection of some of my first projects Data Science projects originally done on Kaggle. Each project has its own directory within this repo, and a brief description can be found in their respective README files.

Projects

Digit Recognition (March 2023):

• This project focuses on building a machine learning model to recognize handwritten digits. It involves exploring the famous MNIST dataset and training various classification algorithms.
• The trained model achieved an accuracy of 99.52% on the testing set, which is a state-of-the-art performance on the MNIST dataset.

House Prices (March 2023):

• In this project, I analyze a dataset of housing prices and develop a regression model to predict the prices of new houses based on their features. I explore various data preprocessing techniques, feature engineering, and model selection.
• My last model ensemble achieved RMSE of 0.119 on the testing dataset, which placed me in the top 4% of the Kaggle House Prices Competition leaderboard.

Superstore Analysis (February 2023):

• In this project, I analyze a fictional superstore dataset to gain insights into its sales, customers, and products. The project involves data cleaning, exploratory data analysis, visualization, and also includes geoanalysis to examine spatial patterns and relationships within the data. The analysis provides valuable insights into sales trends, customer behavior, and geographic factors that contribute to the store's performance.

Matplotlib (December 2022):

• This project provides simple examples of some of the most popular charts from the matplotlib library, including bar charts, line charts, and scatterplots.
• This project serves as a beginner's guide to Matplotlib, demonstrating simple techniques to create clean and visually appealing visuals.

Titanic (November 2022):

• This project is centered around the Titanic dataset, aiming to predict the survival of passengers based on various features. I explore feature engineering, model training, and evaluation techniques to build a predictive model.