• Prince Boateng
• Sandra Ngoing

The project focuses on fraud detection and risk assessment in insurance claims. The primary objectives are:

• Fraud Detection – Identifying fraudulent claims efficiently to minimize financial losses.
• Risk Assessment – Analyzing patterns in incidents and claims to improve risk evaluation and resource allocation. The dataset includes customer demographics, policy details, incident information, and claims data, with a specific fraud_reported variable for fraud detection.

• Fraudulent claims may be more frequent in certain age groups and specific locations.
• Incidents might be more common during certain times, such as rush hours or late evenings.

• Missing Data – Handled using statistical methods.
• Inconsistent Formats – Standardized date/time formats.
• Duplicate Entries – Identified and removed.

• High claim amounts often correlate with higher fraud likelihood.
• Fraud detection models can help insurers adjust policies dynamically.
• Risk-based pricing models can be implemented to charge higher premiums for high-risk customers while offering discounts to low-risk customers.

• Developing a machine learning model to predict fraud can improve detection efficiency.
• Dynamic pricing models can enhance profitability by attracting low-risk customers.
• Insurance claims exhibit randomness, making precise predictions challenging.

Difficulty in importing data into an SQL database, which highlighted the importance of matching data types.

• Dataset: https://www.kaggle.com/datasets
• Database Schema: https://drawsql.app/teams/ironhack-30/diagrams/fraud-detection-in-insurance-claim
• Project Management: https://miro.com/app/board/uXjVLk1X-cM=/
• Presentation: https://docs.google.com/presentation/d/1vKB9qjN7wstwF9-iy4f10cAQeBCf86lfRtjVNStXnLc/edit#slide=id.p