I typically spend 1-2 weeks on personal projects. My bandwidth for long projects is strictly reserved for work. The ideation process is spontaneous, usually when curiosity, a learning opportunity, and scheduling align.
My passion is in solving people and customer problems. I’m very much enjoy navigating through uncertainty and dealing with dense, ambiguous subjects.
People / Talent Analytics
A 3-day lesson plan that combines all the widely used tools (Excel, Tableau, Jupyter Notebook) on the market and introduces students to HR Analytics and its applications.
Talent Migration and Time Series Forecasting
Storytelling through data. A storyboard that explains the past and present talent migration across countries, industries, and skillsets/expertise, and forecasts them for the next 5 years using Holt Exponential Smoothing.
Forecasting Model Demo
Workplace Health & Safety App
A mvp mobile app which projects and recommends workplace safety regulations and return-to-office exigency for employees based on 300+ leading and lagging environmental indicators, job function, and individual health.
Prototype Pitch Draft
Automated Trading System with APIs
A trading bot that trades stocks for you using TD Ameritrade API in Python.
I wrote multiple Classes for trading actions using tda-api. The stock, trading interval, and investment strategy of choice is user-determined. In the example, I used RSI or EMA thresholds (ta for financial ratios) to trigger actions. Process was automated end-to-end with Selenium.
Customer Segmentation via Heurisitics and ML
Customer segmentation and cohort analysis on real retail data using heuristics and k-means clustering.
I performed EDA on 525k records from an UK based eCommerce store, converted prices dynamically based on fx rate at time of purchase (cc), grouped cohorts, and conducted behavioural analysis. My heuristic approach (RFM quartile) was analyzed against results from k-means clustering, revealing 69% congruency.
Customer Segmentation Presentation
Google Cloud Platform ETL and BI
Comprehensive report and three dashboards outlining high-level, ad-hoc, and KPI needs, using tools from Google.
I wrote queries for metrics such as total pageviews, unique pageviews, entrances, exits, time on page, bounces, and more, to identify the disproportionately valuable KPIs which could drive the biggest impact.
Google Merchandise Store Dashboard
Webscraping Three Ways
Using Regex, Selenium, and BeautifulSoup, built data pipelines for three diversely formatted sites to compile financial data (yfinance) into workable dataframes, with emphasis on regex.