Dashboard/Explanation
Every month a dashboard is created for a select set of countries. The dashboard presents a summary of the data situation for a country. The elements of this summary is presented here.
Total number of products
The histogram shows the total number of products for the country in question with the growth on previous month as a percentage.
Nutriscore
The pie-chart shows the percentage of products with and without a (calculated) Nutriscore. For the products that have a Nutriscore, the subdivision by score is shown.
Nutritional Values / Categories
The calculation of the Nutriscore is based on the availability of a category and the nutritional values. The Nutritional values/Categories piechart/doughnut show, which percentage of products have this data available. Read this as green is good and red is bad. The inner ring shows the Categories. The outer ring shows the Nutritional Values. Each colour is subdivided into three shaded, based on the availability of an associated image. The lighter the shade the better, as this implies that the data can be checked against an associated image. We strive to have the +/+-part, i.e. extracted data with selected image to have as large as possible, as this allows us to verify the data. Note that the products in the red light shade area have an image, but the nutritional values have not yet been extracted. The red wedges of both the Nutritional Values doughnut as the Categories piechart must decrease if we want more products with a Nutriscore.Work to do.
NOVA
The pie-chart shows the percentage of products with and without a (calculated) NOVA. For the products that have a NOVA, the subdivision by score is shown.
Ingredients
The NOVA-score is based on the availability of the ingredients. Green is good and red is bad. Each colour is subdivided into three shaded, based on the availability of an associated image. The lighter the shade the better, as this implies that the data can be checked against an associated image. We strive to have the +/+-part, i.e. extracted data with selected image to have as large as possible, as this allows us to verify the data. Note that the products in the red light shade area have an image, but the ingredients have not yet been extracted. Work to do.
Data quality
The data quality pie-chart shows the percentage of products with identified errors. Green is good (no errors detected), red is bad.
Improvement
The improvement graph is a bit special. All changes to database involves only few products. Showing how Nutriscore or NOVA improves over time goes by the percentage or smaller. Such a graph does not provide much information. Another way to look at the evolution is to show how things have change with respect to previous month, and this by month. Each value NOVA, Nutriscore and Data quality can be compared to the overall growth. We would like to see that the NOVA/Nutriscore/Quality values are larger than the overall growth. This means that OFF is able to keep up with the growth and is able to catch up.