The Flemish provinces play an active role in supporting local retail. For years, they have used the Locatus database to gain better insights into retail storefronts in Flanders. The Province of Antwerp wanted to go one step further and looked for additional parameters for all storefronts and, by extension, for the retail area in which they are located.

Pilot project waAI

Under the name “winkelgebieden afbakenen met AI” (waAI, delineating retail areas with AI), the Province of Antwerp launched a pilot project and awarded the assignment to Locatus in collaboration with GIM.
“The proposal and methodology of Locatus and GIM convinced us thanks to their technical expertise combined with the most competitive price,” says Senne Van Hoof, Head of Retail at the Province of Antwerp.

AI model scores appearance

Visual characteristics such as architecture, historic character, window size, and graffiti are important parameters for a storefront’s appearance. The AI model developed by GIM fully automatically analyzes images of thousands of buildings and translates the results into an attractiveness score for each storefront.

In addition to the score, GIM provides supplementary information about facade width. This additional information enriches the Locatus dataset and supports spatial analyses by shopping street or retail area.

From pixels to policy

Together with additional insights from Locatus about the concentration of storefronts in a given area, the Province gains an extra source of information for its retail policy.

This project was made possible with the support of the European Regional Development Fund (ERDF), the Flemish Government (VLAIO) and the Province of Antwerp.