Food retailers’ scanner data provide unprecedented details on local consumption, provided that product identifiers allow a linkage with features of interest, such as nutritional information.

In this paper, we enrich a large retailer dataset with nutritional information extracted from Open Food Facts, completed with the ANSES Ciqual dataset. To compensate for imperfect matching through the bar code, we develop a methodology to efficiently match short textual descriptions. After a preprocessing step to normalize short labels, we resort to fuzzy matching based on several tokenizers (including n-grams) by querying an ElasticSearch customized index and validate candidates echos as matches with a Levenstein edit-distances. The pipeline is composed of several steps successively relaxing constraints to find relevant matching candidates.

We finally develop a similarity based on a word embedding obtained by training a Siamese neural network on bar code matches. This alternative measure is used to evaluate our final matching.

Food retailers’ scanner data provide unprecedented details on local consumption, provided that product identifiers allow a linkage with features of interest, such as nutritional information.

In this paper, we enrich a large retailer dataset with nutritional information extracted from Open Food Facts, completed with the ANSES Ciqual dataset. To compensate for imperfect matching through the bar code, we develop a methodology to efficiently match short textual descriptions. After a preprocessing step to normalize short labels, we resort to fuzzy matching based on several tokenizers (including n-grams) by querying an ElasticSearch customized index and validate candidates echos as matches with a Levenstein edit-distances. The pipeline is composed of several steps successively relaxing constraints to find relevant matching candidates.

We finally develop a similarity based on a word embedding obtained by training a Siamese neural network on bar code matches. This alternative measure is used to evaluate our final matching.

Les résultats principaux de l’étude sont les suivants:

• Si l’étude confirme que l’âge est le facteur de risque prépondérant d’hospitalisation, elle met en évidence d’autres facteurs de risque ;
• La population hospitalisée réside dans des logements plus densément occupés et appartenant plus fréquemment au parc social ;
• Un risque d’hospitalisation décroissant avec le niveau de vie ;
• Une population hospitalisée plus souvent née à l’étranger ;
• La quatrième vague se distingue des vagues précédentes, ce qui pourrait traduire les effets des campagnes de vaccination ;
• Le risque de décéder à l’hôpital reste légèrement moins élevé chez les personnes aux plus hauts revenus ;

Plus de détails ci-dessous ! 👇

We bring together mobile phone and geocoded tax data on the three biggest French cities to shed a new light on segregation that accounts for population flows. Mobility being a key factor to reduce spatial segregation, we build a gravity model on an unprecedent scale to estimate the heterogeneity in travel costs.

Residential segregation represents the acme of segregation. Low-income people spread more than high-income people during the day. Distance plays a key role to limit population flows. Low-income people live in neighbourhoods where the spatial frictions are strongest.