A machine learning project that predicts whether a developer's task will succeed based on daily behavior patterns like coding time, sleep, meetings, and distractions.
- Source: Kaggle – AI Developer Productivity Dataset
- Target Variable:
task_success(Binary: Success or Not)
⚠ Dataset is not included in this repository due to redistribution restrictions.
You can download it directly from the Kaggle link above and place the file ai_dev_productivity.csv in the project folder to run the notebooks.
coding_hourssleep_hoursmeeting_hoursdistractions
Dropped Features:
coffee_intake_mgcognitive_load
Dropped due to multicollinearity to improve model generalization.
| Model | Accuracy | Notes |
|---|---|---|
| Logistic Regression | 0.86 | High precision for successful tasks |
| Decision Tree | 0.96 | Excellent performance, slight overfitting |
| Random Forest | 0.98 | Best overall performance |
- Confusion Matrix:
[[33, 1], [1, 65]] - Classification Report:
- Precision: 0.97 (Class 0), 0.98 (Class 1)
- Recall: 0.97 (Class 0), 0.98 (Class 1)
- F1-score: 0.98 (Overall)
- Cross-Validation Accuracy: ~0.95 (5-fold)
coding_hours– Most impactfulsleep_hoursmeeting_hours
The Random Forest model achieved the highest performance with 98% accuracy, showing that productive coding hours and sufficient sleep are key predictors of task success.
"More sleep and deep work, less chaos — a recipe for coding success."
- Python (Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn)
- Google Colab
Raviha Khan
🔗 LinkedIn
🐙 GitHub
📧 ravihakhan53@gmail.com
📍 Karachi, Pakistan
📱 0332-5214319