Jayana Sarma

I’m a data analyst focused on turning large-scale, messy datasets into clean, decision-ready outputs. I validate demographic and institutional records, automate reporting pipelines, and present insights in formats stakeholders can actually use.

At the Long Beach Dept. of Health & Human Services, I validated U.S. Census ACS 5-year demographic data for 458,491 residents using Python and SQL. I designed and deployed a 5-view interactive dashboard in Python Shiny and Plotly, including CSV/PNG exports and colorblind-accessible styling, and I built geospatial choropleths using GeoJSON and TIGER shapefiles while automating ingestion to reduce reporting errors.

In the Adelante Program at CSU Chico, I merged and validated multi-year institutional datasets across cohorts and academic years in Python, R, and Excel. My analysis surfaced a 7–14% retention gap and a 13% graduation gain for first-generation and underrepresented student cohorts, and I presented results to non-technical program leadership to support retention strategy.

I also bring a building mindset from my data science internship work, including powering statistical analysis with a multi-agent LLM assistant (Azure OpenAI GPT-4) and improving performance with caching logic. Beyond my core analytics work, I’ve developed predictive models like credit card fraud detection, traffic flow forecasting, and stroke risk prediction using tools such as XGBoost, scikit-learn, and Streamlit.