𝐂𝐚𝐧𝐜𝐞𝐫-𝐑𝐢𝐬𝐤-𝐅𝐚𝐜𝐭𝐨𝐫𝐬-𝐄𝐃𝐀-𝐌𝐋

A compact and standardized dataset examining how lifestyle, environmental, and genetic factors influence five common cancer types. Contains 2,000 individual records × 21 features, fully numerical and ready for EDA, dashboarding, and multiclass ML tasks.

---

Why This Dataset?

Explore risk distributions and correlations across lifestyle factors Build visual dashboards for population-level cancer risk Train multiclass models on Cancer_Type with balanced evaluation (macro-F1, accuracy) Practice class imbalance handling and interpretability

---

Targets

Primary target: Cancer_Type ∈ {Lung, Breast, Colon, Prostate, Skin} → Ideal for multiclass classification (use macro-F1, accuracy, confusion matrix).

Optional target:

Risk_Level ∈ {Low, Medium, High} → Derived from Overall_Risk_Score thresholds: Low < 0.35 | 0.35–0.65 = Medium | > 0.65 = High

---

🔍 Analysis Pipeline

Categorical Features:

Analyzed variables like Cancer_Type, Risk_Level, Gender, H_Pylori_Infection, and BRCA_Mutation using bar charts.

Most patients fall into Medium or High risk groups.

H_Pylori_Infection and BRCA_Mutation strongly link to specific cancer types.

Chi-Square Test:

Confirmed strong associations (p < 0.05) between Cancer_Type,

H_Pylori_Infection, and BRCA_Mutation with overall risk level.

Numerical Features:

Variables: Age, BMI, Smoking, Alcohol_Use, Air_Pollution, Physical_Activity, etc.

High-risk patients → higher BMI, Smoking, and Air Pollution.

Low-risk patients → better Physical Activity and Diet Quality.

Correlation & Feature Importance:

Top correlated features with overall risk:

1️⃣ Smoking 2️⃣ Air Pollution 3️⃣ BMI 4️⃣ Age

Data Preparation:

✅ Dropped non-predictive IDs

✅ One-hot encoded categorical variables

✅ Train-test split (80/20)

🤖 Machine Learning Models

Random Forest Classifier:

Accuracy: 96%

Top Predictors: Overall_Risk_Score, Smoking, Air_Pollution, BMI, Age

Logistic Regression:

Accuracy: 85%

✅ Random Forest outperformed across all metrics (F1 ≈ 0.95).

---

📊 Key Insights

Environmental and lifestyle factors dominate cancer risk.

Smoking, Air Pollution, and Obesity are top contributors.

Exercise and Healthy Diet act as strong protective factors.

---