Finovera — AI Stock Portfolio Advisor
An AI-powered app that delivers personalized stock portfolio recommendations based on investor profiles, combining financial data with news sentiment analysis.
Business Context
Individual investors struggle to make informed stock picks without professional tools. Finovera was built to bridge this gap using AI-driven analysis of market data and news sentiment.
Strategic Problem
How to provide personalized, daily stock recommendations by combining financial metrics with real-time news sentiment, tailored to each investor's risk profile?
Data Sources
Historical stock/ETF data (open/close prices, volume, daily variations) via Yahoo Finance API. Press article sentiment scores from NewsAPI using VADER sentiment analysis.
Methodology
Built a full data pipeline fusing market data with sentiment scores. Trained a Random Forest classifier (tested against LogisticRegression, XGBoost, LSTM) to predict daily asset performance. The system selects the 5 highest-probability assets daily for personalized buy recommendations.
Key Results
Functional app with a Streamlit interface and Swift mobile frontend. 44 commits, full ML pipeline from data ingestion to daily recommendations.
Business Impact
End-to-end AI project demonstrating data engineering, ML modeling, NLP sentiment analysis, and product design — from API integration to user-facing app.