PROJECTS

FinVAR: The LLM Version

Developed using LangChain and Chainlit, it leverages and upgrades LLMs as a stepping stone for fiduciary-level stock risk analysis. Built a multi-source data pipeline that analyzes any publicly traded ticker across all market sectors, integrating real-time financial data, news sentiment, and 10-K risk factors from SEC Edgar files, FinnHub, FinancialModelingPrep, and OpenAI APIs. Designed the system to deliver comprehensive risk assessments—from blue-chip stocks to small-cap companies—helping users make informed investment decisions based on live market data. Click on the image to check it out.
Screenshot of FinVAR, the LLM version, a stock investment risk analysis tool, with information on data sources, usage instructions, and risk levels, set against a dark background with colorful icons.

NEWISE_GIS

Created NEWISE_GIS to solve a common problem: finding the right place to live in New England shouldn't require endless Google searches. Engineered an interactive visualization platform that aggregates and displays regional data on income, education, crime, and demographics, helping users make informed relocation decisions through intuitive geographic comparisons. Click on the image to check it out.
Text document titled 'NEWISE_GIS: The New England WISE Geographic Information System' created by Atul Aravind Das, Aaron Huang, Harry Nguyen, Bidipta Roy, describing the platform's purpose and features, with a button labeled 'New England Map' for exploring an interactive map.

FinVAR – Financial Assistant Referee

Created FinVAR to bridge the gap between financial analysis and predictive insights for investors. Traditional tools show historical metrics but lack forward-looking capabilities—FinVAR solves this by combining interactive visualizations of profitability, growth, leverage, and liquidity metrics with a Machine Learning model that predicts future EPS. Engineered a Gradient Boosting algorithm trained on historical financial ratios and integrated the Yahoo Finance API to auto-fetch real-time data, delivering end-to-end analysis from current performance to future projections in a single platform.  Click on the image to check it out.
Screenshot of a digital interface displaying a logo and text for 'FinVAR - Financial Assistant Referee'.

Appropriations Equity Analysis for Massachusetts Senator Edward Markey’s Office

Collaborated with Senator Markey's office to transform Massachusetts funding data into actionable intelligence. Engineered interactive geospatial visualizations using Python, Folium, and OpenStreetMap API to map population and per capita funding across all state municipalities. Built a LookerStudio dashboard for stakeholder presentations and led biweekly strategy sessions to translate data insights into policy recommendations. Click on the image to check it out
Title: Massachusetts Maps. Subheading: Competitive Discretionary Grant Allocation Equity Analysis. Created by Atul Aravind Das, Carlos Garcia, Anthony Huang, John Salloum, and Daniel Strick. List of map topics including city population, per capita funding, median household income, county heat map, race and income cluster map, demographic heat map, and funding heat map.