Page 594 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Overview Tab:

• KPI Metrics: Number of NGOs monitored, fraud alerts generated, projects validated.
• Risk Score Distribution: Bar chart of NGOs by risk level (low, medium, high).

Fraud Detection Tab:
• Anomaly Alerts: Table listing flagged transactions (NGO name, amount, date, risk score).
• Network Visualization: Graph showing connections between suspicious NGOs/directors.

Geospatial Validation Tab:

• Map View: Overlay of reported project locations with satellite imagery.
• Validation Status: Pie chart (verified, unverified, flagged).
Sentiment Analysis Tab:

• Sentiment Trends: Line chart of public sentiment (positive/negative) over time.
• Keyword Cloud: Top fraud-related terms from social media/news.

Technology: Built using Plotly Dash for interactivity.

NGO Fraud Detection Report

• Sections:
1. Summary: Total NGOs analyzed, fraud alerts, and high-risk cases.
2. Financial Anomalies: List of flagged transactions with details (amount, recipient, risk
score).
3. Registration Issues: Duplicate registrations or unverified documents.
4. Project Validation: Discrepancies between reported and actual project locations.
5. Public Sentiment: Summary of negative sentiment and key allegations.
6. Recommendations: Prioritize audits for top 5 high-risk NGOs; investigate specific
transactions.
• Format: PDF exportable from dashboard with NIC branding.

3 Use Case Requirements

REQ-01: The system shall ingest and analyse financial data to detect anomalous transactions
and patterns indicative of fraud.

REQ-02: The system shall identify potentially fraudulent NGOs by detecting duplicate
registrations and suspicious operational or financial profiles.

REQ-03: The system shall monitor NGO compliance with key regulatory requirements, including
reporting obligations.

REQ-04: The system shall leverage location-based data and external imagery to validate the
existence and activities of claimed projects.

REQ-05: The system shall analyze textual data from various sources (social media, news, reports)
to identify potential fraud indicators and public sentiment.

REQ-06: The system shall generate alerts and provide risk scoring to prioritize potential fraud
cases for review.

558