Raw Student Data
Here's a sample of your dataset. The goal is to predict the exam_score.
Feature Engineering Lab
Formula:
Build Your Prediction Model
1. Select Features to Use
2. Model Results
Model Performance (R² Score)
0%
Based on the strongest single predictor
Predictive Power of Each Feature
Data Scientist's Insight 💡
Great attempt! Data scientists often create features that capture an underlying concept.
In this dataset, a powerful feature is study_quality, calculated as (hours_studied * hours_of_sleep) / (cups_of_coffee + 1).
Why it works:
- It rewards focused, well-rested study.
- It penalizes cramming with too much coffee (which leads to diminishing returns).
- This single feature has a much higher correlation with exam scores than any raw feature alone!