Projects
Here's some of my projects.
Interpretable Clustering Pipeline
Developed an end-to-end Python pipeline to solve the 'black box' problem in clustering. This project automates the process from data preprocessing and dimensionality reduction (UMAP/t-SNE) to generating human-readable interpretations (SHAP, decision rules) that explain why data points are grouped together, transforming abstract clusters into actionable insights.
Project Page ↗ GitHub ↗ (opens in a new tab)
Netflix Case Study
Analyzed over 8,800 data points from the Netflix catalog. I successfully identified key trends in viewership, leading to actionable insights on content strategy. My analysis revealed that adding TV Shows during July and December months could increase viewership by up to 15%. The study has equipped decision-makers with data-driven insights, ultimately contributing to a more effective content allocation strategy.
Case Study ↗ (opens in a new tab) Notebook.ipynb ↗ (opens in a new tab)
Aerofit Case Study
My goal in this case study was to identify key characteristics that influence the purchase of specific treadmill models. Through techniques like outlier detection, correlation heatmaps, and probability calculations, I aim to provide actionable business insights. These insights will enable AeroFit to make data-driven recommendations to new customers, potentially increasing sales conversions by an estimated 10-15%.
Unavailable ↗ Notebook.ipynb ↗ (opens in a new tab)
Profit Analysis Dashboard
Interactive "Profit Analysis" dashboard, meticulously designed to offer in-depth insights into a superstore dataset. This intuitive Tableau dashboard provides a comprehensive overview of profit trends over various time frames and categories.
© Made with ❤️ Hitesh Taneja.