How hasanahmeds/capstone-ai-tableau-generator works
This repository builds Tableau dashboards from uploaded tabular data. The user uploads a CSV or Excel file, configures analysis options, runs a staged workflow, then reviews and downloads a generated workbook. The documented pipeline covers validation, profiling, analysis, recommendations, workbook generation, and finalization. The graph preserves those domain stages and the schema, prompt, and resilience layers. Source excerpts are partial, so some documented workflow transitions are shown from workflow.py; optional integrations are marked unverified where their wiring is not shown.
Main entry points: workflow.py Workflow orchestrator · dashboard_analyzer.py AI analysis · app.py Streamlit interface · tableau_workbook_generator.py Tableau workbook generator
Components 18
User experience
- Streamlit interface
app.py
Dashboard pipeline
- Workflow orchestrator
workflow.py - Validate input
workflow.py - Profile dataset
workflow.py - Data processing
data_processor.py - Analyze dataset
workflow.py - Recommend visuals
workflow.py - Generate workbook
workflow.py - Finalize output
workflow.py
Analysis and recommendations
- AI analysis
dashboard_analyzer.py - Analysis prompts
prompt_templates.py - Visualization recommender
visualization_recommender.py
Workbook generation
- Tableau workbook generator
tableau_workbook_generator.py - TWB / TWBX output
Shared models and resilience
- Data and dashboard schemas
schemas.py - Retries and fallbacks
error_handling.py
Other components
- Dashboard user
- LLM provider
Connections 21
- Dashboard user → Streamlit interface: uploads and configures
app.py - Streamlit interface → Workflow orchestrator: runs workflow
app.py - Workflow orchestrator → Validate input: starts with
workflow.py - Validate input → Profile dataset: then profiles
workflow.py - Profile dataset → Analyze dataset: then analyzes
workflow.py - Analyze dataset → Recommend visuals: then recommends
workflow.py - Recommend visuals → Generate workbook: then generates
workflow.py - Generate workbook → Finalize output: then finalizes
workflow.py - Workflow orchestrator → Data processing: uses processing
workflow.py - Workflow orchestrator → AI analysis: runs analysis
workflow.py - Workflow orchestrator → Tableau workbook generator: creates workbook
workflow.py - AI analysis → Analysis prompts: builds prompts
dashboard_analyzer.py - AI analysis → LLM provider: requests analysis
- AI analysis → Retries and fallbacks: uses resilient calls
dashboard_analyzer.py - Workflow orchestrator → Visualization recommender: uses recommendations
- Data processing → Data and dashboard schemas: produces structured data
data_processor.py - Visualization recommender → Data and dashboard schemas: reads and returns specs
visualization_recommender.py - Tableau workbook generator → Data and dashboard schemas: consumes workbook specs
tableau_workbook_generator.py - Tableau workbook generator → TWB / TWBX output: exports
- Finalize output → Streamlit interface: returns results
app.py - Streamlit interface → Dashboard user: shows downloads
app.py