AI Powered Market Intelligence

This project allows you to upload your dataset, generate insights and recommendations, and interactively query your data using AI.

Dataset

• Source: Google Play Store
• Raw Data: googleplaystore.csv
• Cleaned Data: cleaned_googleplaystore.csv
  The cleaning and formatting steps are documented in googleStoreData.ipynb
• Creation of dataframe using Rapid API - Appstore Scrapper API:
  Code implementation is in dataframeThruAPI.ipynb and the dataset created through it is rapidAPIdataframe.csv.
• Combined Dataset:
  We have a combined dataset combined_dataframe.csv. The implementation for combining datasets is in googleStoreData.ipynb; please refer to it.

RAG Pipeline

We use a Retrieval-Augmented Generation (RAG) pipeline with:

• HuggingFace embeddings
• Gemini LLM
• FAISS vector store
• The cleaned dataset: cleaned_googleplaystore.csv

Outputs

• Insights: app_insights.json
• Recommendations: recommendations.json

The main code for generating these outputs is in insights.ipynb and insights.py. For the best experience, run insights.ipynb.

Chatbot

You can query the insights or dataset using a chatbot, also built with the RAG pipeline.
See chatSystem.ipynb (recommended) or chatsystem.py for implementation.