PriorityLens

Customer feedback triage console: your Excel/CSV feedback -> clean -> label (category + sentiment) -> RICE-ranked backlog -> dashboard.

Files

• generator.py - creates a sample Excel file (100 messy rows) i made a sample one, you can use as a template
• pipeline.py - Step 0 load file, Step 1 clean, Step 2 label, Step 3 RICE
• app.py - Streamlit dashboard with a file upload button

Your Excel file

Only one column is required: the feedback text. Name it raw_feedback_text (or feedback, text, comment, review, message). Optional columns: feedback_id, customer_id, source (App Store / Support Ticket / Survey), date. Other source names work too; they just get default reach and confidence values.

How to run

1. Install Python 3.9+ (Windows: tick "Add Python to PATH").
2. Open a terminal in this folder.
3. pip install -r requirements.txt
4. streamlit run app.py
5. Browser opens at http://localhost:8501. Upload your file in the sidebar, or use the sample data.

Without the dashboard: python pipeline.py my_feedback.xlsx

If streamlit is not recognised, use python -m streamlit run app.py.

Notes

• The "LLM" step is simulated with keyword rules (English only). Replace categorize() in pipeline.py to use a real LLM.
• Reach, impact, confidence and effort are editable assumptions in the RICE section of pipeline.py.