Daavi
Verification-first real estate platform restoring trust for the African diaspora.
"Daavi" means "good" in Twi (Ghanaian language) — Named in honor of my late mother
Platform Architecture
A trust-first architecture combining geospatial verification, AI-powered recommendations, and blockchain-ready tokenization infrastructure.
1Problem
The African real estate market suffers from a massive "trust deficit." Diaspora investors lose millions annually to fraud, unverified listings, and lack of transparency.
2Solution
Built a verification-first marketplace using PostGIS geolocation and verified vendor data. The platform includes an AI recommendation engine matching investors to properties based on investment goals.
- →Multi-step vendor verification workflow
- →Geospatial property validation (PostGIS)
- →AI-powered investment matching
3Impact
- ✓Engineered scalable trust architecture using NestJS and TypeORM
- ✓Verification workflow filters out 95% of fraudulent listings
- ✓Seamless cross-border UX for diaspora users in NA/EU
Core Features
Multi-Layer Verification System
Three-tier verification: Document validation, geospatial confirmation (PostGIS), and on-ground inspector network.
AI Investment Matching
Recommends properties based on investment profile (ROI goals, risk tolerance, location preferences) rather than keyword search.
Cross-Border Payment Integration
Secure escrow system supporting multi-currency transactions (USD, EUR, GBP, Local currencies).
Real-Time Market Analytics
Live dashboards showing price trends, neighborhood growth indicators, and investment opportunities.
Technical Deep Dive
Geospatial Verification with PostGIS
Leveraging PostGIS to validate property coordinates against official land registry data and detect location manipulation.
// PostGIS-powered property validation
@Injectable()
export class PropertyVerificationService {
async validateLocation(property: Property): Promise<boolean> {
const query = `
SELECT ST_Contains(
(SELECT boundary FROM verified_zones WHERE zone_id = $1),
ST_SetSRID(ST_Point($2, $3), 4326)
) as is_valid
`;
const result = await this.db.query(query, [
property.zone_id,
property.longitude,
property.latitude
]);
return result.rows[0].is_valid;
}
}Vendor Trust Score Algorithm
Multi-factor trust scoring system considering verification history, user reviews, and completion rate.
// Trust score calculation
interface TrustFactors {
verificationsPassed: number;
avgRating: number;
completedTransactions: number;
accountAge: number; // in days
}
function calculateTrustScore(factors: TrustFactors): number {
const weights = {
verification: 0.4,
rating: 0.3,
transactions: 0.2,
tenure: 0.1
};
const normalized = {
verification: Math.min(factors.verificationsPassed / 10, 1),
rating: factors.avgRating / 5,
transactions: Math.min(factors.completedTransactions / 50, 1),
tenure: Math.min(factors.accountAge / 365, 1)
};
return Object.entries(weights).reduce((score, [key, weight]) =>
score + (normalized[key] * weight * 100), 0
);
}AI-Powered Property Recommendations
Using OpenAI embeddings to match user investment profiles with property characteristics beyond simple filters.
// Semantic property matching
const userPreferences = `
Investment goal: Long-term rental income
Budget: $50k-$150k
Risk tolerance: Moderate
Location preference: Growing urban areas
`;
const embedding = await openai.embeddings.create({
model: "text-embedding-ada-002",
input: userPreferences
});
// Vector similarity search in PostgreSQL with pgvector
const matches = await db.query(`
SELECT *, 1 - (embedding <=> $1) as similarity
FROM properties
WHERE 1 - (embedding <=> $1) > 0.8
ORDER BY similarity DESC
LIMIT 10
`, [embedding.data[0].embedding]);The "Spotify Evolution" Product Strategy
Rather than launching a full-featured platform at once, Daavi follows a phased approach inspired by Spotify's growth—start with core trust, expand to discovery, then scale with advanced features.
- • Vendor onboarding with document verification
- • Property listing with PostGIS validation
- • Basic search and filtering
- • Secure user authentication (JWT)
- • AI-powered recommendation engine
- • Personalized investment dashboards
- • Saved searches and alerts
- • Community reviews and ratings
- • Fractional property ownership (blockchain)
- • Secondary market for property tokens
- • Smart contract-based escrow
- • Liquidity pools for real estate
- • Predictive analytics for property values
- • Neighborhood growth forecasting
- • Investment portfolio optimization
- • Macro-economic trend analysis
Lessons Learned
Trust is a Feature, Not an Afterthought: Initially focused on building a beautiful UI, but user research revealed that diaspora investors care more about verification than aesthetics. Pivoted to make trust visible at every touchpoint.
PostGIS Changed Everything: Adding geospatial intelligence wasn't just a "nice-to-have"—it became the core differentiator. Being able to say "this property has been geo-verified" built instant credibility.
Cultural Context Matters: Payment systems that work in the US don't work in Ghana. Had to integrate local mobile money options (MTN, Vodafone Cash) alongside Stripe. Never assume infrastructure parity.
MVP ≠ Minimum Viable Product: For Daavi, MVP means "Minimum Viable Trust." Without verification, the platform has no value. Better to launch later with robust verification than early with a broken trust model.
What's Next
Q1 2026
- • Beta launch with 50 verified vendors
- • Mobile app (React Native)
- • Virtual property tours (360° video)
Q2-Q3 2026
- • Expand to Nigeria and Kenya markets
- • Partnership with local banks for financing
- • AI recommendation engine v2
2027+
- • Tokenization pilot program
- • Decentralized property registry
- • Pan-African expansion