Student Dropout Early-Warning System using Academic Records

Audience

Who it is for

Teams who want an ML project with a real campus story and fairness discussion.

Counsellor login, top-20 at-risk list, why, and a recorded outreach note.

Branches

Team size: 1–4 students

Problem

Existing vs proposed

Colleges notice dropouts after the student has already left. Spreadsheets do not rank who needs a call this week.

Features from attendance, internals, backlogs, fee delay. Time-based split so you do not train on the future. Ranked list + explanation + action log.

Build

Modules

1. 1. Warehouse

   • Student, course, attendance, marks, fees
   • Anonymised IDs

2. 2. Model

   • Time split, class weights
   • Lift chart for the top decile

3. 3. Explain

   • Per-student reasons
   • Fairness slice by gender/quota if present

4. 4. Counsellor desk

   • Queue, actions, reopen
   • Weekly CSV for HOD

Deliverables

Included with the kit

• Complete source code (ZIP + GitHub-style repo)
• Database schema and sample data / dataset notes
• README.md — how to run, stack, modules, and demo path (Windows + college lab)
• Viva question bank with short answers
• Two mentor reviews: architecture, then final demo