This project analyzes 120+ Scaler mentorship sessions to identify inefficiencies and improve mentor performance, learner satisfaction, and program management.
Using Python, NLP, and visualization dashboards, it integrates AI to automate session summaries, detect sentiment in learner feedback, and flag early warnings — helping Program Managers proactively improve session quality.
Unstructured mentorship sessions and missing feedback led to:
- Inconsistent learner satisfaction
- Low follow-up completion
- Weak documentation and limited data insights
Key Stats:
- Session Completion Rate: 55.8%
- Average Rating: 3.64 / 5
- Missing Feedback: 58.3%
- Follow-up Scheduled: Only 41.7%
- Negative Feedback: 5.8% of total
A 3-phase mentorship model was proposed to improve accountability and engagement:
- AI-generated agendas using learner data.
- Mentor & learner checklists.
- Smart topic recommendations from prior feedback.
✅ Expected Impact: Improved mentor preparedness & reduced cancellations.
- Real-time tracking of engagement and attendance.
- AI-powered note-taking (speech-to-text + summarization).
- Structured conversation flow.
✅ Expected Impact: Better engagement & documentation.
- Automated session summaries emailed to mentors and learners.
- Follow-up scheduling triggers for low sentiment or unresolved goals.
✅ Expected Impact: Continuity, accountability, and improved retention.
Objective: Detect negative learner experiences automatically.
Tech: TextBlob, Seaborn, Pandas
Outputs:
- Classifies feedback as Positive / Neutral / Negative
- Visual correlation between sentiment and ratings
Results:
| Sentiment | % Feedback | Avg Rating |
|---|---|---|
| Neutral | 65.8% | 3.84 |
| Positive | 28.3% | 3.59 |
| Negative | 5.8% | 3.14 |
🪄 Enables early intervention by flagging “Negative” sessions within 24 hours.
Objective: Automate session documentation for mentors & PMs.
Tech: KeyBERT, NLP, Python
Functionality:
- Extracts key terms from learner feedback.
- Auto-generates structured summaries in CSV for dashboards.
Example Output:
- Sneha Reddy conducted a Resume Review session. Limited learner feedback available.
- Session on Career Guidance by Vikram Singh highlighted “motivation, clarity, roadmap.”
✅ Saves 5–7 minutes per session & standardizes weekly reporting.
Dashboard Features:
- Average Rating by Month
- Bottom 5 Mentors by Rating vs Follow-Up
- Feedback Sentiment Analysis
- Early Warning Flags per Mentor
- Key KPIs: Total Sessions, Avg. Rating, Prep %, Follow-up %
Example Metrics:
| Metric | Value |
|---|---|
| Total Sessions | 120 |
| Completed Sessions | 67 |
| Avg. Rating | 3.64 |
| Mentor Prep Done | 56.67% |
| Follow-Up Scheduled | 41.67% |
| Negative Feedback Count | 7 |
Goal: Identify risky sessions early.
Rules used for flagging:
- Low Rating (<3)
- No Mentor Preparation
- No Follow-up
- Negative Feedback
Color Codes:
🟩 All Good | 🟨 Mild Attention | 🟥 Urgent Follow-Up
✅ Enables Program Managers to act proactively using Power BI or Excel dashboard.
| Category | Tools / Libraries |
|---|---|
| Data Processing | Pandas, NumPy |
| Visualization | Matplotlib, Seaborn, Power BI |
| NLP & AI | TextBlob, KeyBERT |
| Automation | Python Scripts, CSV/Excel Export |
| Deployment Ready | Google Colab / Jupyter Notebook |
Data Cleaning → Descriptive Stats → Sentiment Analysis →
Keyword Extraction → AI Summaries → KPI Metrics →
Early Warning Dashboard