Support Ticket Router (Streamlit)

Small ML app that predicts a support queue from ticket text (subject + body).

Screenshot

What it does

• Input: free-form support ticket text (subject + body).
• Output: predicted queue + confidence + top-K alternatives.
• UI: Streamlit app with example buttons and a simple explanation of influential tokens.

Dataset

This project uses the public dataset Tobi-Bueck/customer-support-tickets (CSV) from Hugging Face: https://huggingface.co/datasets/Tobi-Bueck/customer-support-tickets

Quick download used in this repo: https://huggingface.co/datasets/Tobi-Bueck/customer-support-tickets/resolve/main/dataset-tickets-multi-lang-4-20k.csv

The training script expects these columns in data/dataset.csv:

• subject
• body
• queue

Project structure

support-ticket-router/
  app.py
  train.py
  requirements.txt
  README.md
  data/
    dataset.csv
  artifacts/
    model.joblib