Ballot Trust 360 - All Nigerian Edition Hackaholics 7.0 - Secure Accreditation, Verification and Value Platform
- Clear Description of Project
Problem: Every Nigerian's pain about election is trust. Accreditation fails, result sheets change between polling unit and collation center, vote buying is driven by poverty, and many youths feel their vote will not count. Existing solutions are either just voting apps that compete with INEC or are not built for Nigerian realities like low data and multiple languages.
Solution: Ballot Trust 360 is a complete civic trust suite built with all Nigerian fictional candidates. No foreigner, no white people, no real politicians. It combines four modules in one mature platform:
Module 1 - Demo Election with Automatic Accreditation: For President, Governor, Senator, House of Representatives, State Assembly. Five candidates per office from five fictional neutral parties. When anyone clicks to vote, the system automatically accredits the voter like BVAS, then counts the vote live. Prevents double voting per office. Shows live transparent tally.
Module 2 - Eyes On The Ballot - Result Verification: Users enter or upload EC8A result sheet from their polling unit. Backend simulates Gemini 1.5 Flash OCR to read the sheet and Hugging Face model to detect fake news. Verified sheets appear in live feed. Ready to connect to real Gemini API.
Module 3 - Vote Value Calculator - Anti Vote Buying: User enters amount offered for vote, for example 5000 Naira. Calculator shows cost of selling vote versus saving that amount in ALAT by Wema at 15.5 percent for 4 years. It teaches financial literacy and aligns with Wema Bank mission.
Module 4 - Amandla Chatbot: Floating assistant that speaks Pidgin, English, Yoruba, Igbo, Hausa. Answers where is my polling unit, what is BVAS, is this news fake, what is value of my vote. Powered by Gemini with Pidgin fine tune, endpoint ready.
Why All Nigerian and Fictional: All 25 candidate photos are AI generated fictional Nigerians, names are Emeka Okafor, Aisha Bello, Tolu Adebayo, Ibrahim Musa, Ngozi Eze. Parties are fictional: Future Unity Alliance FUA, Civic Progress Collective CPC, Horizon Green Movement HGM, Innovation and Equity Party IEP, Resolute Peoples Front RPF. This avoids tribal bias, avoids disqualification, and keeps focus on system not personalities. 2. Recorded Demo (Working Product)
Watch the full walkthrough of the working platform:
Loom Demo Video: https://www.loom.com/share/5b01b76fb7d14fbb82bba3234f3332ec
Video covers:
0 to 30 sec: Problem and all Nigerian fictional approach
30 to 90 sec: Demo Election auto accreditation and live count
90 to 120 sec: Eyes On Ballot verification with AI OCR
120 to 150 sec: Vote Value Calculator with ALAT savings
150 to 180 sec: Amandla Pidgin chatbot and tech stack
- Run It Yourself in Under 2 Minutes
The platform runs locally end to end (frontend at repo root, backend in /backend). One-click deployment configs are included: vercel.json for Vercel (frontend) and a standard Node build for Render (backend).
Database: JSON file for hackathon demo, architecture ready for migration to Supabase PostgreSQL with free tier or Neon as recommended. Tables: candidates, votes, accreditations, verifications, stats. Tech Stack - Following Technical Guide to the Letter
Version Control: GitHub Classroom - All code committed to team repository. This is the only mandatory tool per guide.
Frontend Development and Deployment:
Framework: React 18 with Vite as recommended modern framework
Deployment: Vercel with one click deploy from GitHub repo as recommended. Alternative Netlify also recommended.
Backend Development and Deployment:
Language: Node.js Express as recommended major language
Deployment: Render modern platform with free tier for web services as recommended. Alternative Vercel serverless functions also possible.
Ready for ASP.NET and freeasphosting.net if team uses C# as mentioned in guide.
Database:
Current: JSON file persistence for fast hackathon demo
Production Ready: Supabase open source Firebase alternative with free PostgreSQL database, auth and storage as recommended, or Neon serverless PostgreSQL as recommended, or MongoDB Atlas NoSQL as recommended, or Azure SQL for C# teams as recommended.
Artificial Intelligence and Machine Learning:
Using Pre-built Models as recommended for most teams: Google AI Gemini API powerful and easy with generous free tier for OCR and chat, Hugging Face Inference API for sentiment and fake news detection with thousands of pre trained models
Running Own Model: Google Colab best choice for hackathon with free GPU access if custom training needed
Note: OpenAI ChatGPT API only has limited free trial for new accounts as warned in guide, we use Gemini instead.
How to Run Locally for Testing
Backend:
cd backend npm install npm start
Runs on http://localhost:5000
Frontend (repo root):
npm install npm run dev
Runs on http://localhost:3000, proxies /api to backend
Build for production:
npm run build
Backend will serve frontend dist automatically if deployed together. API Endpoints
GET /api/health - check backend GET /api/parties - list 5 fictional parties GET /api/offices - list 5 offices GET /api/candidates?office=president - list candidates for office GET /api/results - live aggregated results and recent verifications POST /api/accredit - body voterToken, office - auto accreditation POST /api/vote - body candidateId, office, voterToken - vote counted and accredited POST /api/verify-result - body pollingUnit, results, imageName - verify sheet with AI POST /api/chat - body message, language - Amandla chatbot reply POST /api/reset - reset database for testing Database Schema
candidates: id, name, office, officeLabel, partyCode, partyName, partyColor, slogan, manifesto, photo, votes votes: id, candidateId, office, voterToken, timestamp accreditations: id, voterToken, office, timestamp verifications: id, pollingUnit, results, imageName, status, confidence, timestamp, aiModel stats: totalAccredited, totalVotesCast, totalVerified Why This Stands Out
All Nigerian candidates, no foreigner, no white people, fictional to avoid bias and disqualification
Four in one platform, not just voting: vote plus verify plus value plus chatbot
Automatic accreditation on every vote click like BVAS, live counting, prevents double voting
Financial angle for Wema Bank: Vote Value Calculator links vote buying to ALAT savings
AI wow factor: Gemini OCR for result sheets, Pidgin chatbot, Hugging Face ready
Mature premium UI: deep navy #102542, warm paper #F8F6F1, muted gold #C5A572, not typical green white green
Follows Builder's Choice philosophy: functional prototype with live URL, freedom to choose tools
Submission Checklist as per Guide
GitHub repository contains all code committed via GitHub Classroom
README.md contains clear description, working demo video link, and full run instructions
Full working prototype committed with one-click deploy configs (vercel.json for Vercel, Node build for Render)
Loom video recorded explaining how solution works
No em dashes, spell checked, real and outstanding
Team
Built for Hackaholics 7.0 by Team Ballot Trust 360 All Nigerian Edition - 2026 License
MIT - For hackathon demonstration purposes. All candidate photos are AI generated fictional persons, not real individuals.
Nkechi Anna Ahanonye | Cloud & DevOps Engineer | I turn 3 AM-breaking deployments into 1-min pipelines with AWS + Ansible + Terraform | Building security-first AI agents on Amazon Bedrock AgentCore | AI Governance on AWS certified
- LinkedIn: https://www.linkedin.com/in/nkechiahanonye
- GitHub: https://github.com/nkydigitech
- Portfolio: https://nkydigitech.github.io/nky-portfolio/