Reading the Signs: What 5 Years of Real EstateData in Lille Tell Us

"AN INSIDE LOOK AT THE PATTERNS,

OUTLIERS, AND HIDDEN STRUCTURES "

Behind every price tag lies a story—of location, size, timing, and human decisions.

Before building models, we first had to understand what Lille’s housing market was

telling us.

Using five years of notarial records from the official DVF dataset, we analyzed over

44,000 apartment transactions in Lille from 2020 to 2024. Each one held essential

signals: where the property was, how big it was, when it sold, and for how much.

This was our foundation.

Neighbourhood Trends

We grouped Lille into 10 zones based on transaction patterns. These clusters reflect the character and pricing of real neighbourhoods:

• Lille Centre (Cluster 6) – Highest prices per square meter, with compact

surfaces.

• Wazemmes (Cluster 1) – A lively district with good value for medium-sized

apartments.

• Hellemmes (Cluster 2) – Lower prices and larger living spaces.

  
  

This allowed us to compare pricing patterns across the city more meaningfully,

beyond just street names or postal codes.

Apartment Attributes

Focusing only on apartment sales, here’s what we saw:

• Most apartments ranged between 30 and 100 m²

• On average, units had 3 to 4 rooms

• The majority were priced between €100,000 and €400,000

• Luxury properties, mainly in central zones, reached up to €1.2 million

  
  

We also calculated total lot size per unit to reflect shared or subdivided property

surfaces. Where this wasn’t available, we treated it as a standalone unit.

In addition to interior space, we examined lot information when available. Some

properties, especially in shared buildings- included subdivided lots, which we

aggregated to understand their total surface profile.

A small portion of properties exceeded €1.2 million, mostly located in Lille

Centre. While rare, these transactions reflect real demand for luxury housing and

were kept in our analysis to preserve market realism.

Why It Matters and What’s Next

Great models start with great data—but even more important is the clarity of

analysis, which reveals how cities evolve.

This study uncovered how location, size, and timing influence property value in

Lille. From the compact, high-priced streets of Vieux-Lille to the emerging

opportunities in areas like Hellemmes, the numbers reveal real stories.

"A well-prepared dataset is not just a file-it’s a structured view of the city’s

story." — Team Amzil AI

We’re now translating these findings into an AI-powered valuation tool—designed

to understand how surface area, geography, and timing combine to shape prices.

In our next release, we’ll show how the system uses this knowledge to assist real

estate decision.