ForeSee

Inspiration

ForeSee was inspired by a common experience many students and professionals share: facing a dataset and wishing there were an easier way to analyze it. The idea grew from the question, “Wouldn’t it be great if there were a tool that could automatically interpret data and explain it in simple terms?” That thought became the foundation for creating a platform that enables anyone to perform data analysis without needing to code or understand complex models.

What it does

ForeSee simplifies machine learning reporting. You can upload a CSV file to get an automatic overview, select a target variable, and instantly receive a generated report with predictions, metrics, and clear explanations. Using SHAP values, Foresee explains why the model made its predictions, in plain English, and provides a downloadable PDF file.

How we built it

Front-End:

React 19 with Vite.
TailwindCSS for a modern design system.
React Router for navigation
AOS for animations
ESLint for code linting

Back-End:

Python 3.11+ as the core backend language.
Flask3.0 for building a lightweight, high-performance REST API.
Flask-CORS for cross-origin requests

AI & ML:

Google Gemini 2.0/2.5 (via google-generativeai): Target variable recommendations, natural language insights generation.
scikit-learn 1.5.0 (Logistic Regression, Decision Tree)
XGBoost 2.1.0 (Gradient Boosting)
SHAP 0.44.0 (Model explainability)
pandas 2.2.0 & NumPy 1.26.0

Data Platform:

Snowflake (Data Warehouse) snowflake-connector-python 3.12.0, snowflake-snowpark-python 1.39.1

Report Generation:

ReportLab 4.0.7 for PDF Generation
Matplotlib 3.8.0 for charts and visualizations

Challenges we ran into

Connecting multiple systems while keeping the response times as low as possible was a challenge.
Balancing visual accuracy and performance when generating PDF charts.
Implementing a complete end-to-end pipeline, from dataset to upload to explainable model reports within 24 hours demanded focus and prioritarization.

Accomplishments that we're proud of

We built a full pipeline that takes raw data, runs ML models, explains them with SHAP, and produces a PDF report
Implemented a working prototype of intellingent agents communicating through innovative protocols
We designed a responsive frontend using React, TailwindCSS, and Figma for prototyping.
Balanced frontend, backend, and ML components in 24 hours

What we learned

We learned how to coordinate across multiple roles: frontend, backend, design, and machine learning
We gained hands-on experience combining several technologies, from Flask to TailwindCSS
We deepened our understanding of SHAP and LIME, and how explainability can make machine learning more human-centered
Building a complex system under a tight hackathon deadline taught us to prioritize features with the most value quickly.

What's next for ForeSee

Public deployment
User accounts to allow users to sign in, save datasets, and track previous analyses securely
Move beyond static PDFs to dynamic and shareable dashboards
Integrate with could platforms to handle larger datasets
Enable teams to annotate and comment directly on reports or shared analysis