Developed a Financial Fraud Detection system during my internship at Infotact Solutions. The project combined Python-based machine learning with Power BI dashboards to identify and visualize fraudulent activities in financial transactions.

🔹 Approach:

• Cleaned and preprocessed 100K+ credit card transactions in Python.

• Built a Random Forest Classifier to predict fraudulent cases.

• Evaluated the model using Confusion Matrix, Precision-Recall, and ROC Curve.

🔹 Key Insights from Dashboard:

• Fraud Rate: 1% of total transactions (1K frauds out of 100K).

• Top High-Risk Cities: Chicago, San Diego, Dallas.

• Monthly Fraud Trends: Peaks in January and July.

• Fraud by Transaction Type: Refunds and Purchases equally impacted.

• Fraud vs Legitimate Distribution clearly showing class imbalance.

🔹 Impact:
Enabled risk teams to quickly analyze fraud hotspots, trends, and transaction anomalies, making fraud detection more interpretable for decision-makers.