SCORECHAIN – Wallet Risk Scoring

This project calculates a risk score (0–1000) for Ethereum wallets using their on-chain transaction history.

---

🔍 Problem Statement

You are given wallet addresses. The objective is to:

1. Fetch transactions from Etherscan.
2. Create features based on transaction behavior.
3. Apply both rule-based and model-based scoring strategies.

---

📦 Data Source

• Transactions fetched using Etherscan API
• Feature set includes:
  • Total transactions
  • ETH sent / received
  • Number of failed transactions

---

🧠 Rule-Based Risk Scoring

Rule	Penalty
tx_count < 5	-200
failed_tx > 2	-150
total_out > total_in	-100

Score Formula:

score = 1000 - penalties

---

🤖 Model-Based Risk Scoring (KMeans)

Features Used:

• tx_count
• total_in
• total_out
• failed_tx

Steps:

1. Normalize features
2. Cluster wallets using KMeans (2 clusters)
3. Score based on distance to “safe” cluster centroid

Visualization:
Wallets are projected to 2D using PCA and colored by cluster.

---

📁 Input Format

An Excel file with wallet addresses:

wallet_id
0xabc...
0x123...

---

📤 Output Files

• wallet_risk_scores_combined.csv: Contains both rule-based and model-based scores
• kmeans_cluster_visualization.png: Cluster visualization

---

🚀 How to Run

Install Dependencies

pip install pandas requests scikit-learn openpyxl matplotlib

Replace your Etherscan API Key

Edit wallet_risk_scoring_combined.py:

ETHERSCAN_API_KEY = "YOUR_KEY_HERE"

Run the Script

python wallet_risk_scoring_combined.py

---

🧪 Normalization

Normalization was applied only for model-based scoring using MinMaxScaler.
Rule-based scoring relies on fixed thresholds and does not require scaling.

---

Author

Vanga Jai Prakash
Creator of ScoreChain2 – Wallet Risk Scoring (+ LLM add-on)

• Email: vangajaiprakash@gmail.com
• GitHub: Vanga-Jai-Prakash