Home›Projects›Fraud Detection›Sandbox
OverviewDocumentationMethodologyUse Cases🧪 Sandbox

Fraud Detection — Transaction Scanner

Test how a fraud analytics workflow can score transactions, identify suspicious patterns, prioritise investigations, and estimate loss-prevention value. The demo uses demo data by default and can adapt to uploaded CSVs with common transaction column names.

Privacy: Data is processed in-browser for this static portfolio demo.
Output dashboard

Dataset preview comes first so users can confirm the data loaded and mapped correctly before reviewing the fraud visuals.

Risk distribution
Rule breakdown
Amount exposure by risk tier
Category risk profile
Detection insights
Ask the Fraud Assistant
No dataset loaded
Load transaction data and score it to get fraud investigation guidance.
💳
Load transactions to begin
Use the demo dataset or upload a CSV. The scanner will map columns, calculate risk scores, display dataset-grounded filters, create dynamic visuals, and produce an investigation queue.
How this sandbox can be tested by users
  1. Load sample data and confirm the preview, row count, field list and risk KPIs update.
  2. Filter by country, category or channel and confirm charts and queue recalculate.
  3. Move the minimum risk score slider to test analyst workload versus fraud recall.
  4. Upload a compatible CSV and confirm column mapping, dynamic filters and scored output adapt to the uploaded schema.
  5. Export the investigation queue for downstream review, compliance documentation or case-management handoff.