Paris properties tokenization with smart contracts based on live smart city data
DocPropertiesTokensData sourcesOntologyAppraiseCityPulse is a working smart-city data platform for Paris. It streams real measurements from the city's own infrastructure (road-traffic loops, air-quality stations, bike-share docks), combines them with every official property sale and the city's amenities, and organises it all into a live digital model of the city. On top of that model it values homes, and then issues those homes as tokenized assets on a blockchain: a title deed, 1,000 tradable shares, an on-chain valuation oracle and a marketplace that settles instantly.
It is built the way enterprise smart-city platforms (such as Palantir Foundry) are built: connectors land raw data, distributed pipelines refine it, an ontology turns tables into connected real-world objects, and applications such as this dashboard run on that ontology. All city and property data is real open data; the investors and share trades on the chain are simulated to demonstrate the market.
Property sales
official DVF records, 2021–2025
Road-traffic sensors
induction loops, hourly
Bike-share stations
Vélib' live docks
City amenities
metro, RER, schools, parks, shops
Sensor observations
time-stamped, in the ontology
Tokenized properties
78 share trades settled
Smart-city technology
sensing → data platform → ontology → decisions
A smart city is only as smart as its ability to connect what its sensors measure with the decisions people make. CityPulse covers that whole chain, end to end:
Sense
IoT feeds & open data
Land
raw, append-only zone
Refine
PySpark pipelines
Model
city ontology
Decide
valuation & apps
Live urban sensing
1,756 road-traffic induction loops report flow and occupancy every hour; air-quality stations report PM2.5, NO₂ and the European Air Quality Index per district; 1,519 bike-share stations report dock availability in real time. Collectors poll these feeds on a schedule.
Lakehouse data platform
Raw readings are stored untouched as an audit trail, then PySpark pipelines clean, type and deduplicate them and run geospatial joins: every sale, sensor and amenity is placed in its district, and every home is linked to its nearest metro, RER station, school and park.
City ontology (digital twin)
The refined data becomes a graph of real-world objects (14 object types: districts, sensors, sales, amenities, tokens, investors) and the links between them. The ontology is derived from the pipelines, never edited by hand, so it always reflects the latest data.
Open standards
The graph exports to JSON-LD using W3C SOSA/SSN, the sensing vocabulary behind NGSI-LD and SAREF4City, plus schema.org for places and sales. That makes it interoperable with standard smart-city platforms rather than a closed silo.
Location intelligence
Each home is described by what the city measures around it: walking distance to transit, schools and parks within 500 m, the district's road-traffic load and its air quality. These urban signals are exactly what drives its value.
Explainable valuation
A hedonic model prices every home from those city signals, with a per-driver explanation. It was tested honestly on 28,958 sales from 2025 that it never saw during training: median error 15%, in line with what a public sales register without floor or condition data allows.
Real-world asset tokenization
property → deed → shares → oracle → market
Tokenization turns ownership of a real asset into a digital token on a blockchain. For real estate this means a property can be split into small, freely tradable shares, settled in seconds instead of weeks, with every transfer recorded in a public, tamper-proof ledger. CityPulse implements the full lifecycle with five smart contracts written in Solidity. The demo runs them on a local Ethereum-compatible chain; the same contracts deploy unchanged to any EVM network:
Appraise
city-data valuation
Mint
title deed NFT
Fractionalize
1,000 shares
Anchor
valuation oracle
Trade
instant settlement
Index
back into the ontology
Digital title deed
Each property is minted as a unique token (ERC-721) whose identifier is derived from the official land-registry parcel id, so the blockchain and the city ontology refer to the same real asset.
Fractional ownership
The deed is locked in escrow and 1,000 fungible shares are issued against it. The deed can only leave escrow if someone gathers 100% of the shares. A 500k€ apartment becomes 1,000 shares of 500€.
Valuation oracle
The city-data valuation is anchored on-chain: a cryptographic fingerprint of each appraisal, its fair value and the model that produced it. Personal data stays off-chain (GDPR); only its proof is public, so anyone can verify a price.
Atomic marketplace
Investors list and buy shares at a fixed price. Payment, in a euro stablecoin (a demo stand-in for a regulated one), and shares move in a single transaction or not at all: no counterparty risk and no settlement delay.
Chain ↔ city data, one model
An indexer reads the contracts' events back into the data platform, where they become ontology objects (tokens, investors, holdings, offers, trades). One graph walk goes from the original sale to the appraisal, the deed and today's cap table.
Why it matters
For a city developer, tokenization means fractional access to prime real estate, faster capital recycling, liquidity for traditionally illiquid assets and a transparent registry, all on the same data platform that already runs the city.
What each page shows
click a card to open the page
Properties
The investment view: a map and sortable table of 2,213 real Paris homes with asking price, model fair value and how far apart they are. Filter by district, type, budget, rooms and distance to transit, schools and parks. Click a home for its full dossier: value drivers in plain language, nearby amenities with walking times, live traffic and air quality, price trend, opportunity score, 3-year projections and, if tokenized, its on-chain status.
Tokens
The blockchain explorer: every tokenized property with its deed, share supply, last traded price and market value; the cap table of who owns how many shares; and a live feed of mints, appraisals, offers and trades.
Data sources
The data catalog: each feed's provider, licence, update frequency, coverage and freshness, plus exactly which ontology objects and valuation signals it feeds. Click any sensor on the map to see its reading history.
Ontology
The city's digital model: how many objects and links of each type it holds, a lookup that resolves any object and follows its links, and a district view combining sales, sensors, amenities, air quality and hourly traffic.
Appraise
Price any home: choose a district, type, surface and rooms, and get a fair value with the contribution of each city signal, next to a price map and the district's price trend.