AGSPE — Adani Group Strategic Prediction Engine

Available Languages

Last Run Information

Workflow	Last Run	Status
Data Ingestion (every 4hrs)
Model Validation (weekly)

An open-source automated intelligence platform 😝 developed under the broader research and transparency ethos associated with the Cockroach Janta Party movement — focused on ingesting public data about the Adani Group ecosystem, validating information through a multi-tier cross-source protocol, and generating probabilistic predictions on future business moves with confidence scores and timelines.

Inspired by the satirical anti-corruption and transparency oriented political culture associated with the Cockroach Janta Party, AGSPE emphasizes open data analysis, source transparency, public accountability, and verifiable intelligence workflows built entirely on publicly available information.

Yeh Website pakka chalta hai 😅😝

Live website: https://cockroachjanathaparty-agspe.vercel.app/

Vision

AGSPE is designed as a public-interest analytical platform that combines:

Open-source intelligence (OSINT) Cross-source validation Bias-aware scoring systems Public financial analysis Transparent probabilistic forecasting

The project reflects the broader Cockroach Janta Party philosophy of:

decentralized public scrutiny, anti-corruption transparency, open civic technology, and accessible investigative tooling.

AGSPE does not claim insider access, privileged information, or political affiliation with any corporate or governmental entity. All analysis is derived from publicly accessible datasets, regulatory filings, news archives, and algorithmic inference models.

Features

Predictive Analytics Dashboard — Hybrid prediction engine combining financial indicators (35%), political alignment scoring (35%), and pattern recognition (30%) to forecast probable business moves with confidence intervals
Multi-Source Validation Engine — Three-tier source trust weighting (Tier 1 = 0.9, Tier 2 = 0.7, Tier 3 = 0.3) with cross-verification rules, contradiction detection, and pro-group source capping
Financial Intelligence — Dual-currency (INR/USD) display of 10 listed Adani Group companies with market cap, debt-to-equity heatmap, P/E ratios, and sortable tables
Lobbying & Political Tracker — Global lobbying spend across 5 countries, electoral bond donation tracker, law firm engagement mapping
Intelligence Feed — Filterable news feed with source tier badges, validation score bars, bias detection warnings, and keyword search
Acquisition History — 20 major acquisitions catalogued with valuations, status tracking, sector filtering, and year-range filtering
Bias Detection — Automatic flagging of pro-group sources, Tier 3 claims without Tier 1 corroboration, and contradiction alerts between Tier 1 sources
Daily Data Pipeline — Vercel Cron Jobs fetch fresh data from public sources (Reuters RSS, BBC RSS, SEC EDGAR) every day at 6 AM UTC, run validation and prediction pipelines, and update the dashboard
Zero-Config Deployment — Deploy to Vercel in one click; works immediately with built-in mock data as fallback — no API keys required

Core Principles

Principle	Description
Transparency	Every prediction includes validation metadata, source weighting, and confidence scoring
Public Data Only	No private databases, leaks, or unauthorized access
Open Verification	Claims are cross-checked across independent media and regulatory filings
Bias Awareness	Source weighting and contradiction detection are visible to users
Civic Technology	Built as a public-interest research platform using open-source infrastructure
Explainability	Prediction logic and scoring models are documented and auditable

Tech Stack

Layer	Technology
Frontend	Next.js 16, React 19, Tailwind CSS 4, Shadcn/UI, Recharts
API / Backend	Next.js API Routes (TypeScript), Vercel Serverless Functions
Validation Engine	TypeScript port of Python multi-tier verification algorithm
Prediction Engine	TypeScript hybrid model (Financial + Political + Pattern)
Data Fetching	`fetch()`-based RSS/API crawlers (serverless-compatible)
Cron Jobs	Vercel Cron (daily at 6 AM UTC)
Full-Stack (Optional)	Python 3.11 + FastAPI + Celery + PostgreSQL + Redis
CI/CD	GitHub Actions (3 workflows)

Why This Exists

AGSPE emerged from the idea that:

large corporate ecosystems increasingly influence public infrastructure, energy, logistics, finance, and policy,
public understanding of these systems often depends on fragmented reporting,
and open-source analytical tooling can help researchers, journalists, students, and citizens better understand systemic corporate behavior patterns.

The platform therefore attempts to:

aggregate dispersed public information,
validate claims through weighted-source verification,
identify recurring strategic patterns,
and present findings in an auditable and transparent manner.

Research Orientation

The project draws inspiration from:

OSINT methodologies,
investigative journalism workflows,
financial risk analysis systems,
civic-tech transparency initiatives,
and decentralized open research communities.

AGSPE is intentionally structured so that:

all scoring logic is inspectable,
all datasets can be audited,
all source weighting rules are visible,
and all predictions remain probabilistic rather than deterministic.

Quick Start

Option 1: Deploy to Vercel (Recommended — One Click)

Click the button above or import the repo on vercel.com/new
Vercel auto-detects Next.js — just click Deploy
Your dashboard is live at https://your-project.vercel.app
Cron jobs are automatically configured via vercel.json

No environment variables required. The app works immediately with built-in mock data. The cron job will attempt to fetch live data from public RSS feeds and APIs; if those are unavailable, it gracefully falls back to the mock dataset.

Option 2: Local Development

# Clone the repository
git clone https://github.com/cockroachparty/cockroachjanathaparty-agspe.git
cd cockroachjanathaparty-agspe

# Install dependencies
npm install

# Start development server
npm run dev

Open http://localhost:3000 — the dashboard works immediately. All API routes (/api/predictions, /api/financials, etc.) return mock data by default.

To test the cron refresh pipeline locally:

curl http://localhost:3000/api/cron/refresh

Option 3: Full Stack with Docker Compose (Advanced)

For the full Python backend with PostgreSQL and Redis:

cp .env.example .env
# Edit .env with your API keys (optional)
docker-compose up --build

This starts:

Next.js Frontend on http://localhost:3000
FastAPI Backend on http://localhost:8000 (Swagger docs at /docs)
PostgreSQL on localhost:5432
Redis on localhost:6379
Celery Worker for async data ingestion tasks

Vercel Deployment Details

How It Works

AGSPE runs entirely on Vercel as a single Next.js application:

Frontend — React components rendered by Next.js, fetching data from API routes
API Routes — Serverless functions at /api/* that serve data from an in-memory cache
Data Pipeline — Vercel Cron triggers /api/cron/refresh daily, which:
- Fetches latest news from Reuters RSS, BBC RSS, The Hindu RSS
- Fetches regulatory filings from SEC EDGAR public API
- Runs all articles through the Multi-Source Validation Engine
- Generates new predictions using the Hybrid Prediction Engine
- Updates the in-memory cache with validated, scored data
Graceful Fallback — If external sources are unavailable, all data falls back to a comprehensive mock dataset

Cron Schedule

Schedule	Endpoint	Description
Daily at 6 AM UTC	`/api/cron/refresh`	Fetches fresh data, runs validation & prediction pipelines

The cron is configured in vercel.json and activates automatically on deployment.

Optional Environment Variables

Variable	Purpose
`CRON_SECRET`	If set, the `/api/cron/refresh` endpoint verifies the `Authorization: Bearer <CRON_SECRET>` header
`NEWS_API_KEYS`	Comma-separated News API keys for enhanced news fetching

These are entirely optional — the app works without them.

Vercel Configuration

The vercel.json file configures:

{
  "crons": [{
    "path": "/api/cron/refresh",
    "schedule": "0 6 * * *"
  }],
  "functions": {
    "src/app/api/cron/refresh/route.ts": {
      "maxDuration": 60
    }
  },
  "headers": [{
    "source": "/api/(.*)",
    "headers": [
      { "key": "Cache-Control", "value": "s-maxage=3600, stale-while-revalidate=86400" }
    ]
  }]
}

Cron runs daily at 6 AM UTC
The refresh function has a 60-second timeout (max for Vercel Hobby plan)
API responses are cached for 1 hour with 24-hour stale-while-revalidate

API Endpoints

Endpoint	Method	Description
`/api`	GET	System info (name, version, status, last refresh)
`/api/predictions`	GET	Strategic predictions with confidence scores
`/api/financials`	GET	Financial data for 10 Adani Group companies
`/api/news`	GET	Validated news articles with bias scores
`/api/lobbying`	GET	Lobbying records and electoral bond data
`/api/acquisitions`	GET	Acquisition history with valuations
`/api/health`	GET	System health status and data freshness
`/api/cron/refresh`	GET	Cron-triggered data refresh pipeline

All endpoints return JSON with the structure: { "data": [...], "source": "live" | "mock", "lastRefresh": "ISO timestamp" }

Project Structure

├── .github/workflows/
│   ├── data-ingestion-cron.yaml      # Scheduled data pipeline (every 4hrs)
│   ├── model-validation.yaml         # Weekly model backtesting
│   └── security-scan.yaml            # Security audits on push
├── backend/                           # Python FastAPI (optional full-stack)
│   ├── app/
│   │   ├── api/routes.py             # FastAPI endpoints
│   │   ├── crawlers/                 # Python web crawlers
│   │   ├── models/                   # Python engine code
│   │   ├── data/                     # Seed data & source tiers
│   │   ├── utils/                    # Currency, text, audit
│   │   └── config.py                 # Backend config
│   ├── tests/                        # Python test suite
│   ├── Dockerfile
│   └── requirements.txt
├── docs/
│   ├── data_dictionary.md            # Data entity documentation
│   └── model_logic.md                # Prediction model documentation
├── src/
│   ├── app/
│   │   ├── api/                      # Next.js API Routes (Vercel backend)
│   │   │   ├── route.ts              # Root API endpoint
│   │   │   ├── predictions/route.ts  # Predictions API
│   │   │   ├── financials/route.ts   # Financials API
│   │   │   ├── news/route.ts         # News API
│   │   │   ├── lobbying/route.ts     # Lobbying API
│   │   │   ├── acquisitions/route.ts # Acquisitions API
│   │   │   ├── health/route.ts       # Health check
│   │   │   └── cron/refresh/route.ts # Cron refresh pipeline
│   │   ├── layout.tsx                # Root layout
│   │   ├── page.tsx                  # Main dashboard page
│   │   └── globals.css               # Global styles
│   ├── components/
│   │   ├── dashboard/
│   │   │   ├── PredictionDashboard.tsx
│   │   │   ├── FinancialsTab.tsx
│   │   │   ├── LobbyingTab.tsx
│   │   │   ├── IntelligenceFeed.tsx
│   │   │   └── AcquisitionsTab.tsx
│   │   └── ui/                       # Shadcn/UI components
│   ├── lib/
│   │   ├── engine/                   # TypeScript engine code
│   │   │   ├── types.ts              # Shared types & config
│   │   │   ├── validation.ts         # Multi-Source Verification Engine
│   │   │   ├── predictor.ts          # Hybrid Prediction Engine
│   │   │   ├── bias-detector.ts      # Bias detection
│   │   │   └── data-fetcher.ts       # Serverless data fetcher
│   │   ├── data/
│   │   │   └── store.ts              # In-memory data cache
│   │   ├── mock-data.ts             # Mock dataset (fallback)
│   │   ├── utils.ts                  # Utility functions
│   │   └── db.ts                     # Database client
│   └── hooks/                        # Custom React hooks
├── public/
│   └── logo.svg                      # AGSPE logo
├── docker-compose.yml                # Full-stack Docker config
├── Dockerfile.frontend               # Frontend Docker
├── vercel.json                       # Vercel cron & config
├── .env.example                      # Environment variables
├── next.config.ts                    # Next.js configuration
├── package.json
├── LICENSE                           # MIT License
└── README.md

Validation Engine

The Multi-Source Verification Algorithm is the core of AGSPE's reliability:

Source Trust Tiers

Tier	Weight	Examples	Behavior
Tier 1	0.9	Reuters, AP, BBC, Bloomberg, Financial Times, OCCRP	Highest trust; can corroborate Tier 3 claims
Tier 2	0.7	The Hindu, Scroll.in, Business Standard	Independent Indian media; moderate trust
Tier 3	0.3	NDTV, Times Now, Republic TV, Adani Group PR	Requires Tier 1 confirmation within 24hrs

Cross-Verification Rules

Tier 3 claims without Tier 1 corroboration within 24 hours get their validation score capped at 0.3 and are flagged "Requires Verification"
Contradicting Tier 1 sources trigger a manual review flag and pause confidence scoring
Pro-group-only sources (all sources are pro-Adani) get scores capped at 0.2
Single source penalty — any claim backed by only one source suffers a 20% score reduction

Output

Field	Type	Description
`validation_score`	Float (0.0–1.0)	Weighted confidence based on source tiers
`bias_risk_level`	String	"Low" / "Medium" / "High"
`flags`	Array	Specific validation flags raised

Prediction Model

The Hybrid Prediction Engine generates forecasts using a three-signal system:

Signal Weights

Signal	Weight	Description
Financial Indicators	35%	Debt-to-equity, capex, volatility, revenue growth
Political Alignment	35%	Proximity to government policy, relevance scoring
Pattern Recognition	30%	Known acquisition cycles, policy correlations

Known Patterns

Pattern	Typical Duration	Base Confidence
Standard Acquisition Cycle	18 months	70%
Green Energy Expansion	24 months	75%
Port & Infrastructure Play	12 months	80%
Distressed Asset Acquisition	15 months	65%

Policy Correlations

Trigger	Predicted Action	Base Probability
BJP infrastructure push	Adani port/airport bid win	82%
Renewable target raised	Adani Green expansion/JV	78%
Defense procurement policy	Adani Defense contract	65%
Housing/infra policy boost	Cement capacity expansion	72%

Data Pipeline

Daily Refresh Flow

Vercel Cron (6 AM UTC)
    │
    ▼
/api/cron/refresh
    │
    ├── Fetch Reuters RSS → Parse articles
    ├── Fetch BBC RSS → Parse articles
    ├── Fetch The Hindu RSS → Parse articles
    ├── Fetch SEC EDGAR API → Parse filings
    │
    ▼
Validation Engine
    │
    ├── Apply tier weights (0.9 / 0.7 / 0.3)
    ├── Cross-verification checks
    ├── Contradiction detection
    ├── Pro-group bias capping
    └── Generate validation_score + bias_risk_level
    │
    ▼
Prediction Engine
    │
    ├── Calculate financial signals
    ├── Score political alignment
    ├── Match acquisition/policy patterns
    └── Generate predictions with confidence + timeline
    │
    ▼
Update Cache
    │
    ├── Store validated articles
    ├── Store fresh predictions
    ├── Store financial data
    └── Update lastRefresh timestamp
    │
    ▼
Dashboard Updated

Graceful Degradation

If external data sources are unreachable:

The data fetcher logs the error and continues
The store retains the previous cache (or mock data on first run)
Predictions are generated from available data
The dashboard always displays data — never shows a blank state
The /api/health endpoint reports the data source as "mock" when using fallback data

GitHub Actions Workflows

Workflow	Trigger	Description
`data-ingestion-cron.yaml`	Every 4 hours / manual	Runs Python crawlers + validation (full-stack mode)
`model-validation.yaml`	Weekly Monday / manual	Runs test suite + backtests predictions
`security-scan.yaml`	Push to `main`/`develop`	Dependency audits, credential scanning, GDPR check

Environment Variables

See .env.example for the full list. None are required for Vercel deployment.

Variable	Required	Description
`CRON_SECRET`	No	Optional bearer token to secure the cron endpoint
`NEWS_API_KEYS`	No	Comma-separated News API keys for enhanced fetching
`DATABASE_URL`	Full stack only	PostgreSQL connection string
`REDIS_URL`	Full stack only	Redis connection for caching
`SECRET_KEY`	Full stack only	Application secret key
`PROXY_ROTATOR_URL`	Full stack only	Proxy service for web scraping

Security & Ethics

No PII collection — The system does not collect or store any personal information
Public data only — All data sources are publicly available records, filings, and news
GDPR/DPDP India compliant — Automated compliance checks in CI/CD pipeline
Rate limiting — All crawlers respect rate limits (2-second delay, max 30 requests/minute)
Bias transparency — Every data point displays its validation score, source tier, and bias risk level
Not financial advice — This tool is for informational and research purposes only

Disclaimer

This project is for informational, educational, and research purposes only.

AGSPE does not provide:

investment advice,
legal advice,
insider information,
political endorsements,
or financial recommendations.

All predictions are probabilistic estimates derived from publicly available information and historical pattern analysis. Users should independently verify all findings before making business, legal, political, or financial decisions.

The project reflects open-source civic-tech and transparency ideals associated with the Cockroach Janta Party ecosystem, but operates as an independent research and software initiative.

License

This project is licensed under the IIT License [custom Cockroach License]— see the LICENSE file for details.

Name		Name	Last commit message	Last commit date
Latest commit History 115 Commits
.github		.github
backend		backend
db		db
docs		docs
prisma		prisma
public		public
src		src
.env.example		.env.example
Caddyfile		Caddyfile
Dockerfile.frontend		Dockerfile.frontend
LICENSE		LICENSE
README.md		README.md
README_hindi.md		README_hindi.md
bun.lock		bun.lock
components.json		components.json
docker-compose.yml		docker-compose.yml
eslint.config.mjs		eslint.config.mjs
next-env.d.ts		next-env.d.ts
next.config.ts		next.config.ts
package.json		package.json
postcss.config.mjs		postcss.config.mjs
tailwind.config.ts		tailwind.config.ts
tsconfig.json		tsconfig.json
vercel.json		vercel.json

Folders and files

Latest commit

History

Repository files navigation

AGSPE — Adani Group Strategic Prediction Engine

Available Languages

Last Run Information

Vision

Features

Core Principles

Tech Stack

Why This Exists

Research Orientation

Quick Start

Option 1: Deploy to Vercel (Recommended — One Click)

Option 2: Local Development

Option 3: Full Stack with Docker Compose (Advanced)

Vercel Deployment Details

How It Works

Cron Schedule

Optional Environment Variables

Vercel Configuration

API Endpoints

Project Structure

Validation Engine

Source Trust Tiers

Cross-Verification Rules

Output

Prediction Model

Signal Weights

Known Patterns

Policy Correlations

Data Pipeline

Daily Refresh Flow

Graceful Degradation

GitHub Actions Workflows

Environment Variables

Security & Ethics

Disclaimer

License

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Contributors

Uh oh!

Languages