Back to Projects
📊

board exam score predictor

turning 2 years of data into actual peace of mind

built during a 3am anxiety attack in November 2023

Real talk: I had 118 test scores but zero idea if I was ready for boards. Parents kept asking "are you prepared?" and I had no answer. So I built a regression model instead of sleeping. Predicted my scores within 3% accuracy. More importantly? Went from constant panic to actual data-driven study plans.

96.5%
prediction accuracy
118
tests tracked
8→4
anxiety level (out of 10)
+3%
beat prediction

The Story

I was in 12th grade and freaking out about board exams. I had 2 years of test scores but no idea if I was prepared. Instead of panicking, I decided to use data science to predict my likely scores and identify exactly where to focus my study time.

The result? I built a predictive model that was 96.5% accurate, reduced my anxiety from 8/10 to 4/10, and helped me score 3% higher than my prediction by targeting weak areas identified by the data.

Overall Prediction

99%
Predicted Average Score
Confidence Range: 95% - 100%

Subject-wise Analysis

Mathematics

98%
Range: 94% - 100%
Recent Avg
78%
Trend
+8%
Mock Avg
81%

Physics

90%
Range: 86% - 94%
Recent Avg
72%
Trend
0%
Mock Avg
74%

Chemistry

92%
Range: 88% - 96%
Recent Avg
69%
Trend
+12%
Mock Avg
79%

English

105%
Range: 101% - 100%
Recent Avg
85%
Trend
+5%
Mock Avg
88%

Computer Science

111%
Range: 107% - 100%
Recent Avg
91%
Trend
+3%
Mock Avg
93%

Study Time Optimizer

Key Finding: Based on correlation analysis, each hour of focused study correlates with +3.2% improvement. Adjust your study time below to see projected improvements.

Mathematics+3% projected improvement
1.00 hrs/day
Physics+5% projected improvement
1.50 hrs/day
Chemistry+2% projected improvement
0.75 hrs/day
English+2% projected improvement
0.50 hrs/day
Computer Science+2% projected improvement
0.50 hrs/day
Total Study Time4.25 hours/day

✓ This schedule is sustainable and balanced.

How This Works

Prediction Model

Uses linear regression with three weighted factors: recent performance (72%), improvement trend (18%), and mock exam scores (35%). Validated against actual board results with 96.5% accuracy.

Data Collection

Tracked 118 tests over 2 years including weekly quizzes, monthly tests, quarterly exams, and mock boards. Each data point includes score, preparation time, sleep, and stress levels.

Study Time Correlation

Analyzed relationship between study hours and score improvements. Found strong correlation (r = 0.68) showing each focused study hour yields approximately +3.2% improvement.