Top 5 Data Science Capstone Project Ideas That Will Impress Employers and Sharpen Your Skills

📗 Chapter 5: Disease Risk Prediction (e.g., Diabetes, Heart Disease)

Build a Health Risk Classifier Using Machine Learning

🧠 Introduction

Healthcare is one of the most impactful domains for data science. By leveraging patient data and machine learning, we can predict disease risk early — improving diagnosis, resource planning, and patient outcomes.

In this capstone project, you’ll build a binary classification model to predict the likelihood of a patient developing a disease such as diabetes or heart disease using structured medical data.

You’ll learn to:

• Load and clean healthcare datasets
• Explore and visualize patterns in risk factors
• Train classification models (Logistic Regression, Random Forest, etc.)
• Evaluate models using sensitivity, precision, ROC-AUC
• Apply feature importance and explainability techniques

Let’s build a tool that helps save lives with data.

🎯 Objective

Goal: Predict the risk of a patient having a chronic disease (e.g., diabetes or heart disease) using clinical attributes such as BMI, blood pressure, glucose, cholesterol, etc.

Dataset: We’ll use the PIMA Indian Diabetes Dataset from Kaggle.

🔗 PIMA Diabetes Dataset

📊 Step 1: Load the Dataset

python

import pandas as pd

df = pd.read_csv('diabetes.csv')

df.head()

🧹 Step 2: Exploratory Data Analysis (EDA)

python

df.info()

df.describe()

Check Class Balance

python

df['Outcome'].value_counts(normalize=True)

Outcome = 1 → Diabetic
Outcome = 0 → Non-diabetic

Visualize with Histograms

python

import seaborn as sns

import matplotlib.pyplot as plt

sns.histplot(df['Glucose'], kde=True)

sns.boxplot(data=df, x='Outcome', y='BMI')

🧼 Step 3: Handle Missing and Invalid Values

Replace invalid zeros with NaNs:

python

cols_with_zero = ['Glucose', 'BloodPressure', 'SkinThickness', 'Insulin', 'BMI']

df[cols_with_zero] = df[cols_with_zero].replace(0, pd.NA)

Impute with median:

python

df.fillna(df.median(), inplace=True)

🏗️ Step 4: Feature Scaling & Splitting

python

from sklearn.model_selection import train_test_split

from sklearn.preprocessing import StandardScaler

X = df.drop('Outcome', axis=1)

y = df['Outcome']

scaler = StandardScaler()

X_scaled = scaler.fit_transform(X)

X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, stratify=y, test_size=0.2, random_state=42)

🤖 Step 5: Train Classification Models

Logistic Regression

python

from sklearn.linear_model import LogisticRegression

lr = LogisticRegression()

lr.fit(X_train, y_train)

Random Forest

python

from sklearn.ensemble import RandomForestClassifier

rf = RandomForestClassifier(n_estimators=100)

rf.fit(X_train, y_train)

📈 Step 6: Model Evaluation

python

from sklearn.metrics import accuracy_score, precision_score, recall_score, roc_auc_score, confusion_matrix, classification_report

y_pred = rf.predict(X_test)

print("Accuracy:", accuracy_score(y_test, y_pred))

print("Precision:", precision_score(y_test, y_pred))

print("Recall:", recall_score(y_test, y_pred))

print("ROC AUC:", roc_auc_score(y_test, rf.predict_proba(X_test)[:, 1]))

Confusion Matrix

python

sns.heatmap(confusion_matrix(y_test, y_pred), annot=True, fmt='d', cmap='Blues')

plt.xlabel("Predicted")

plt.ylabel("Actual")

plt.title("Confusion Matrix")

📊 Step 7: Feature Importance

python

import numpy as np

features = X.columns

importances = rf.feature_importances_

feat_imp = pd.Series(importances, index=features).sort_values(ascending=True)

feat_imp.plot(kind='barh', title='Feature Importance')

plt.show()

🧠 Step 8: Explainability with SHAP (Optional)

python

!pip install shap

import shap

explainer = shap.Explainer(rf, X_train)

shap_values = explainer(X_test)

shap.summary_plot(shap_values, X_test, plot_type='bar')

🩺 Step 9: Real-World Application

Your model can be used for:

• Preventive healthcare alerts in clinics
• Patient triage systems
• Personal health monitoring apps

You can also build a Streamlit app for patients to input their data and check risk level.

📋 Summary Table