How yshukla2024/ojt-ai-projects works
Project(s) and Their Purpose
• Chatbot-Streamlit – A simple web UI (Streamlit) exposing a conversational AI model.
• ML-Models – A notebook for training or experimenting with ML models on restaurant data.
• Python-EDA – A Jupyter notebook-driven exploratory data analysis on the Zomato dataset.
• PowerBI-Dashboard – A BI dashboard (Power BI) visualizing superstore sales.
Pinpoint Main Components & Boundaries
• Data Sources
– restaurant_data.csv (ML-Models)
– Zomato-data-.csv (Python-EDA)
– superstore_sales_full.csv (PowerBI-Dashboard)
• Compute/Processing
– Jupyter Notebooks (project.ipynb, Data-Analysis.ipynb)
– Python code in app.py
• Model Artifacts
– Serialized model files (if created) or in-memory model in notebooks
• Presentation / UI
– Streamlit app for chat interface
– Power BI service/dashboard
Interactions & Data Flows
• Data Ingestion → EDA Notebook
• EDA insights feed feature engineering into ML-Models
• Trained model from ML-Models → deployed in Chatbot-Streamlit
• PowerBI Dashboard reads CSV periodically or on demand
Architectural Patterns & Principles
• Separation of Concerns
– Each subproject has its own directory, config, and data.
• Data Pipeline Pattern
• Layered Approach
– Data layer (raw CSVs)
– Processing layer (notebooks / scripts)
– Serving layer (Streamlit, PowerBI)