I’m building a browser-based dashboard that pulls system-usage data from multiple APIs, normalises it on the fly, then turns the stream into clear, actionable visuals. The data sources you’ll wire up are satellite imagery and structured responses from user-input surveys. After cleaning and merging the feeds, the platform should surface insights in real-time through interactive heat maps; other chart types can remain optional for later phases. A core requirement is an embedded anomaly engine able to flag: • Data discrepancies • Unusual user activity • System errors When an anomaly appears, the system must automatically compile a brief HTML/PDF report and archive it, tagging both the raw and processed records so future queries remain traceable. Key pieces of the build • API connectors for imagery and survey endpoints • Normalisation layer that unifies spatial, temporal, and categorical fields • Real-time datastore (PostgreSQL/PostGIS, TimescaleDB, or comparable) • Heat-map rendering in the dashboard (D3.js, Deck.gl, or similar) • Anomaly-detection logic—rule-based to start, but structured so machine-learning models can be dropped in later • Report generator with templated summaries and visual snapshots • Dockerised deployment scripts and concise setup docs Acceptance criteria 1. A demo environment reachable via URL that ingests sample imagery + survey payloads, updates without manual refresh, and highlights injected anomalies within five seconds. 2. At least three stored reports proving the automatic generation pipeline. 3. Clean, commented code in a Git repo with a README explaining local setup, environment variables, and test data seeding.