How shriyansh24/info_viz_iterration works
Project Type & Purpose
• This is a self‑contained Python “full‐stack” dashboard application built with Streamlit.
High‐Level Architecture Components
• User Interface (Streamlit)
• Application Orchestrator (app.py)
• UI Sections (components/overview.py, insights.py, prediction.py)
• Data Layer
• Input files (attached_assets/*.csv, .xlsx)
• Data processing scripts (data_processor.py + utils/*.py)
• Model Layer
• Trainer (models/trainer.py)
• Predictor (models/predictor.py)
• External Services & Libraries
• Streamlit framework
• Pandas, NumPy (data manipulation)
• Scikit‑learn (machine learning)
• Plotly (visualization)
• OpenpyXL (Excel I/O)
Relationships & Data Flow
• Browser ⟶ Streamlit Frontend (app.py renders pages)
• Frontend calls Data Layer for ingestion/validation
• Processed data ⟶ ML Predictor for on‑the‑fly predictions
• Optionally, processed data ⟶ ML Trainer for batch retraining
• Model outputs & metrics ⟶ UI Sections for visualization & recommendations
• Model performance metrics logged back for Model Monitoring
Architectural Patterns & Principles
• Layered architecture:
• Presentation (Streamlit UI)
• Application (app.py controller)
• Domain (data processing & ML modules)
• Data (file storage)
• Modularization: each major feature (overview, prediction, insights) lives under components/
• Separation of concerns: data_utils vs. model_utils vs. UI components