Cognitive Security & Disinformation Analysis Dashboard

Synthetic, educational dashboard demonstrating how cognitive security teams could triage fictional content for manipulation, misinformation, and coordinated influence signals. All data and outputs are synthetic and non-political.

At a Glance

• FastAPI backend with modular analysis engines and strong input validation.
• Linguistic, misinformation, influence, and risk fusion heuristics to illustrate cognitive security workflows.
• Static dashboard (HTML/JS/CSS) visualizing scores, clusters, and recommended actions.
• Reporting utilities for PDF export and synthetic visuals.
• Safety first: offline, synthetic-only processing—no real individuals or political content.

Quickstart

1. Create a virtual environment and install dependencies:

   python -m venv .venv && source .venv/bin/activate
   pip install -r requirements.txt

2. Run the API locally:

   uvicorn backend.main:app --reload

3. Or start via Docker Compose:

   docker compose up --build

4. Open frontend/index.html in a browser. Ensure the API base URL matches your host/port (default http://localhost:8000).

Architecture

flowchart TD
    UI[Frontend Dashboard] -->|POST /analyze_text| API((FastAPI API))
    UI -->|POST /risk_score| API
    subgraph Engines
      TA[Text Analyzer]
      LD[Linguistic Detector]
      MD[Misinfo Detector]
      ID[Influence Detector]
      RE[Risk Engine]
    end
    API --> TA
    TA --> LD
    TA --> MD
    TA --> ID
    LD --> RE
    MD --> RE
    ID --> RE
    API --> RE

Loading

API Overview

Key endpoints (full reference in docs/API.md):

• POST /analyze_text – Analyze synthetic text payloads and return embeddings plus linguistic & misinformation signals.
• POST /analyze_image – Base64 image input with simulated OCR to text.
• POST /analyze_url – Offline-simulated URL fetch with fictional content.
• POST /risk_score – Fuse multiple posts into a cognitive risk score with influence clustering.
• GET /dashboard – Dashboard readiness summary.
• GET /health – Health probe.

Safety & Scope

• All analyses are synthetic and fictional; no real individuals or political content.
• URL handling is simulated; no external network calls are performed.
• Outputs are for educational demos of cognitive security concepts only.

Project Layout

backend/
  api/           # FastAPI routers
  engines/       # Analysis engines
  utils/         # Shared utilities (logging, OCR stub, embeddings, visuals, pdf)
frontend/        # Static dashboard assets
models/          # Placeholder for synthetic artifacts
logs/            # Runtime logs
Dockerfile       # Production-ready container image
docker-compose.yml # Optional compose stack for local runs
Makefile         # Common developer commands

Development

• Install dev tooling: pip install -r requirements-dev.txt
• Format: black, isort
• Lint: ruff check .
• Tests: pytest

Running Tests & Checks

make lint
make test

Developer Toolbox

• make format to auto-format (black, isort) and apply ruff check fixes.
• .editorconfig to standardize editors/IDEs.
• Docker image includes en_core_web_sm so spaCy loads out of the box.

Code Quality

• Type hints across modules.
• Centralized logging with rotation (backend/utils/logger.py).
• Input validation via Pydantic models at the API boundary.
• Synthetic-only data paths to prevent real-world targeting or data collection.

Reporting & Governance

• PDF export available via backend/utils/pdf_export.py.
• Visual reports (heatmaps, timelines) written to reports/ (auto-created when visuals run).
• Security posture and responsible-use guidelines in SECURITY.md.
• Architecture & design rationale in docs/ARCHITECTURE.md.

License

Released under the MIT License. See LICENSE.