I'm Divya, a data analyst who cleans messy datasets, builds dashboards people actually use, and explains what the numbers mean in plain language. SQL, Python and Power BI are my main tools.
I'm an MCA graduate (2026) with a foundation in statistics and data analysis, and hands-on experience turning spreadsheets and databases into dashboards that hold up under real questions from stakeholders.
My work spans the full pipeline: cleaning and validating messy data, writing SQL to answer specific business questions, building interactive Power BI and Excel dashboards, and explaining findings to people who don't want to read a query. I've applied this across restaurant, retail, telecom and gaming datasets.
Currently based in Bangalore and looking for data analyst roles where I can keep doing this — closer to real business decisions.
Structured and validated telecom churn data in MySQL, resolved data-quality issues in the TotalCharges field, then built SQL logic to calculate churn KPIs across contract types, payment methods and tenure. The Power BI dashboard tracks churn rate, revenue lost, and customer segments in real time.
Built an interactive Power BI dashboard analyzing sales, profit margin and discount impact across regions, categories and products. Regional revenue analysis and profit-vs-discount scatter views surfaced where discounting was actually eroding margin.
An interactive Streamlit app for sales performance and inventory optimization — restocking recommendations, overstock risk flags, and demand trends, built entirely in Python.
Statistical analysis of a mobile game's gate-placement experiment. Used chi-square testing on retention data to evaluate whether moving the game's first paywall actually changed player engagement.
An end-to-end ML project for a growth use case: predicting which customers on a quick-commerce app will stop ordering in the next 30 days. Built a SQL feature-engineering layer (CTEs, window functions) on top of raw order data, then trained and compared Logistic Regression, Random Forest and HistGradientBoosting models — evaluated on ROC-AUC, PR-AUC and F1, with feature importance surfaced for the growth team to act on.