Deepa Sijoy
BI Analyst & data analytics consultant — SQL, Power BI and Python, applied to real business problems in banking, education and NGO operations. 10+ years of operations experience, plus a growing set of independent analytics projects over the last two years.
Projects
An AI-powered platform helping trailing spouses and expats restart their careers abroad — resume parsing, transferable-skill analysis, and live job matching against a skill taxonomy. Selected for DoraHacks 2.0 (1 of 500 teams from 6,000+ applicants) and launched on Product Hunt.
Directed and validated the build with AI-assisted development (Claude Code) — product scoping, data model, and QA were mine; implementation was AI-assisted.
End-to-end churn prediction and analytics solution: Logistic Regression, Random Forest and Gradient Boosting models identify at-risk customers, feeding a six-page interactive Power BI dashboard covering revenue-at-risk, churn drivers, and retention recommendations.
A Logistic Regression model predicting loan eligibility for a mock housing-finance scenario, deployed as a Flask web app with MySQL-backed user registration and login, and instant eligibility results.
Consolidated multi-sheet teacher feedback from a STEM training program into a relational data model (FactFeedback + Master Teacher table), then built a multi-page Power BI dashboard with drill-through by school/zone, plus VADER sentiment analysis on open-text responses.
Portfolio version uses anonymized, representative data — the original was built on proprietary NGO data; methodology and dashboard design are unchanged.
Feature engineering and exploratory analysis on a social media dataset, feeding a Random Forest model that predicts engagement. Identified like-to-impression ratio, impressions, and share ratio as the strongest drivers, with recommendations for content strategy.