Attendance Tracking from Microsoft Teams Report

Summary

I built this to stop matching Microsoft Teams attendance by hand. It cleans up participant names, fuzzy-matches them against the official roster, and shows the result in a Streamlit app.

Problem

Tracking attendance from Microsoft Teams used to mean lining up an attendance report against a participant database by hand.

What made it painful:

• The names in a Teams report never match the official roster
  • Names come out abbreviated
  • Some rows carry only a first and last name
  • Others carry Teams labels like (Unverified), (External), (guest)
  • Capitalization, punctuation and formatting all drift
• There was a lot of it
  • 10+ classes, all handled the same way
  • Several sessions per class, each with dozens of participants
• Doing it by hand is slow
  • It takes hours, and mistakes are almost guaranteed
  • Auditing the result is hard, and keeping it consistent is harder

So attendance became hours of work nobody should do by hand.

---

Solution

I wrote a Python script that: reads the Teams report, cleans the names, matches them against the roster, and totals who showed up.

graph LR
    A[Upload Participant DB] --> B[Upload Teams Report]
    B --> C{Process}
    C --> D[Normalize Names]
    D --> E[Fuzzy Match]
    E --> F[Set Status]
    F --> G[Download Results]

Workflow:

1. Upload the participant database
2. Upload the Microsoft Teams attendance report
3. The system takes it from there:
   • Cleans and standardizes participant names
   • Applies fuzzy matching with a threshold you can tune
   • Determines attendance status
   • Aggregates attendance data across sessions
4. You download the finished sheet.

It runs as a Streamlit web application, so the people who actually need it never touch Python.

---

Impact

Work that took an evening now finishes in minutes, and the numbers are consistent across classes. I have reused it across 10+ classes without changing the workflow:

• ⏱️ Time back
  • Name matching runs on its own now
• 📉 Fewer mistakes
  • Name matching is standardized and consistent
• 📊 Clean, structured attendance data
  • Easier to analyze, report and audit
• 🚀 Reusable
  • Runs across multiple classes with no workflow changes
• 💼 Ready for real use
  • Fits Edu-Tech programs, training programs and stakeholder reporting

---

Tech Stack

• Python
• Pandas for data processing and aggregation
• RapidFuzz for fuzzy string matching
• Regex (re) for text preprocessing and normalization
• Streamlit for the web interface
• Microsoft Teams attendance report as the data source

---

Extras (Optional)

• The fuzzy matching threshold is configurable, so it adapts to how messy the data is
• Parsing, matching, and reporting live in separate modules, so I could retune the threshold without touching the rest
• What I’d add next:
  • CSV / Excel export
  • Analytics dashboard
  • Direct database or LMS integration
  • Automated attendance reporting
• The threshold is the part I would not trust blindly. I still set it by eye on each new dataset, and a truly messy name list will fool it at any threshold.