FinTrend

The Challenge

Retail investors and boutique trading firms lacked access to institutional grade predictive models that could synthesize real time financial news sentiment with historical price movements. Existing tools were either too complex to deploy or relied solely on historical numerical data without context.

Our Solution

We built FinTrend, a full stack platform that brings advanced quantitative analysis to the web. It pairs 24 distinct deep learning models (LSTM, GRU, CNN, RNN) with natural language processing to deliver holistic stock forecasts, enabling users to benchmark how market news impacts predictive accuracy.

Key Features

• ✓Real time dashboard for historical data visualization
• ✓Side by side model comparison and performance benchmarking
• ✓NLP sentiment analysis integration using the FNSPID dataset
• ✓FastAPI backend serving 24 diverse neural network architectures

Tech Stack