AI-DataScience-Lab

Context

A full-stack forecasting tool built to explore the complete data science pipeline end to end, from preprocessing to model deployment.

Problem

Most personal data-science demos stop at a notebook. This one needed to be a real deployed app: upload a CSV, get back cleaned data, a fitted model, and a plain-language summary.

Approach

The app supports CSV uploads, visualizes trends, fits a scikit-learn regression model, and summarizes the dataset using OpenAI's GPT-3.5 API.

Architecture

Backend (Flask) hosted on Microsoft Azure App Service; frontend deployed via GitHub Pages (React).

Results

Version 1.0.0 (Beta) currently supports linear regression, with future plans to add polynomial, ridge, and time-series models for enhanced forecasting.

Stack

Flask · Azure App Service · React · scikit-learn · pandas · OpenAI GPT-3.5 API