namR attributes
The namR attributes are the original items of data in our data store, produced at national level.

They follow a very structured production process, each step of which requires specific expertise.

Process for constructing an attribute’s values

Variety of algorithms and business rules
used for inference

Variety of algorithms

The latest computer vision algorithms are used by the namR teams to create original and complete datasets.

Geomatics algorithms

The manipulation of GPS coordinates in space, precise both at the scale of the building and for the entire country, is a specific characteristic of namR teams.

Business rules

Some attributes come from proprietary business rules defined by our internal experts or in cooperation with business partners. For example, they are used to calculate solar potential, energy consumption or the potential of energy renovations.

Spatial machine learning

Our algorithms are capable of automatically detecting similarities between buildings according to spatial proximity and triangulation to infer the values ​​of certain attributes. Our similarity descriptors are contained in an internal database called the Feature StoR.

Natural Language Processing

Natural Language Processing algorithms are also used to output text elements automatically.

Address and building repositories
Quality assessment
Scalable and optimised architecture