Projects

Here you will find a curated selection of personal and professional projects that demonstrate my ability to deliver value across a variety of domains. These projects span corporate analytics initiatives, deep learning and NLP experiments, machine learning models, data visualisation, web scraping and more.

Corporate Projects

In my professional roles at Amazon and Ensemble Education Innovations, I have delivered impactful data science and engineering projects that optimized supply chain operations, marketing strategies and partner experience.

• Cube Forecasting for Middle Mile Tactical Optimization: Designed hybrid forecasting models using Prophet and XGBoost to predict package volumes and oversize share across routes. Achieved <3 % error and reduced overcapacity costs by ~15 % through proactive capacity planning.
• ITTO Forecasting Data Pipeline: Built Redshift‑SQL and Python ETL pipelines to supply regular, reliable data for forecasting tools across Mexico, Australia and Brazil, reducing manual processing by 80 % and ensuring data consistency across geographies.
• THANOS Tool Development & Data Pipeline: Contributed to the design and build of Thanos, an automated network optimisation and scenario simulation tool for the Middle East & North Africa supply chain. I developed end‑to‑end ETL workflows and QuickSight dashboards to power the tool, enabling self‑service scenario planning, reducing data preparation time by 60 % and accelerating data‑driven decision making.
• BxGy Promotional Strategy Dashboard: Constructed SQL pipelines and QuickSight dashboards for BxGy (Buy × Get y) campaigns in the Hardlines category, saving 12 hours per month and improving promotional decisions.
• Customer Comment Sentiment Analysis: Performed sentiment analysis on customer reviews using Python NLP libraries such as NLTK and TextBlob, identifying friction points and boosting value‑added services by 55 %.
• Report Automation for Affordability & SPEED Programs: Automated flash reports and WBR/MBR decks via SQL pipelines aligned to Andes rules, maintaining 90–99 % on‑time delivery and increasing stakeholder agility.
• VAS Demand Coverage Metric: Redesigned SQL logic and built pincode‑level views to monitor service coverage in the VAS category, increasing coverage to 52 % and closing gaps.
• Insights & Tactical Callouts: Delivered weekly insight decks and category deep dives for VAS and SPEED programs, sharing 20+ insights that drove corrective actions across fan, sports, cycle and kitchen categories.
• Delivery Partner Allocation Optimization: Developed a Python optimization algorithm and SQL pipeline to align delivery partner schedules with preferences, improving preference alignment by 25 % and operational efficiency.
• Leadership Dashboard for Japan: Built interactive QuickSight dashboards from sentiment‑tagged survey feedback to surface critical issues and drive policy changes for partner wellbeing in Japan.

Deep Learning & NLP

My deep learning and NLP projects explore unstructured data and advanced models. Highlights include an NLP analysis of airline company tweets, a dog breed identification system using convolutional networks and a Netflix recommendation system.

Machine Learning

• Predicting census income using diverse ML techniques
• Winning the space race with data science
• Pokémon classification with decision trees
• Heart disease prediction
• Analysing live shopping on Facebook with K‑means clustering
• Customer churn prediction
• Bank loan prediction using classification
• Salary classification from census data via logistic regression

Data Analysis & Visualization

• Healthcare case study: exploring strategies for a wellness tech company
• YouTube data analysis
• Time series analysis of stock market data
• Geospatial analysis case study on Zomato
• Sales insights: data analysis and visualization
• Sports analytics: comprehensive IPL dataset analysis (2008–2021)
• Covid‑19 data exploration and visualization
• Banking dataset exploratory data analysis

Web Scraping

I built scripts to extract and visualise stock market data from the web.

Power BI Dashboards

• Sales insights dashboard
• Sales total vs. average dashboard
• TEDx Talk challenge dashboard
• Nobel Prize winners dashboard

Python Projects

Implemented a classic Tic‑Tac‑Toe game using Python.