Monitoring added products - en

Introduction

The guidelines presented here are based on my work on products added on the OpenFoodFacts-page for the Netherlands.

If you feel like doing this for your country use this url: https://nl.openfoodfacts.org/entry-date/2022-09-27. Adapt the country prefix to your situation. And edit the date for the one you want to use.

Before starting your process, answer the questions that Robotoff proposes.

Products with no data

Quite often one sees products where no data at all has been entered, all fields are empty and there are no images. These products are mostly created by third party apps. They are due to a quirk in the api-calls. Whenever an api interrogates the database with a barcode number and the barcode is not present, the corresponding product is created.

We have no way to ascertain whether these empty products do really exist, or are even food products. Sometimes non-product barcodes can be seen, i.e. barcode lengths other than 8,12 or 13 numbers.

Having these empty products in the database has no value and thus they can be deleted. A user can always add them later again. You might double check by scrolling to the bottom of the page, which will show the status Empty.

Products without images

For these products we can only check the fields that have been filled in and infer the data of other fields. Do not worry to make a mistake here, these can always be corrected. You might have a look at the following fields:

Category

• empty: it can be derived from the product name (if present). Or maybe you recognise the product name.

Products with images

Main language

• filled: check wether the main language is correct. It should be the same as the language found on the ingredients or nutritional values. If not change the main language and transpose all the corresponding data. Save and continue editing.

Multiple languages

• filled: do the other languages really exist. If yes check their data. If not remove the images, product name and ingredients for the non-existing language.

Product name

• filled: does it correspond to what can be found on the images?
• not filled: try to find it on the images, or maybe you recognise it.

Front image

• filled: does it show the front of the product?
• not filled: is a front image available? Assign that one and crop to highlight the image.

Packaging

• not filled: you might recognise the packaging (glass, tin, plastic)

Brand

• filled: is it correct?
• not filled: can you find it on the images?

Shops

• filled: is it a show that exists?
• not filled: for shop brands you might be able to enter the shop name.

Ingredients image

• not filled: add it if available.

Ingredients

• filled: you might want to check it against the image (if available). Try to correct any typos. If it is bad you might want to restart.
• not filled: has the ingredients image been properly cropped, so that it shows only the ingredients? Use the extract button the extract the ingredients, edit if needed. Try to do this also for languages and scripts you do not know. If the extraction does not work you can do it by hand. Sometimes the ingredients image only shows the ingredients partially. Then you might be able to fill the gaps, or add only the ingredients you do see.

Traces

• filled: check. be sure that these are the traces and not the allergens found on the ingredients.
• not filled: can you find it on an image?

Nutritional table image

• not filled: add and crop it if available. Try to use the colour correction to get a better image

Nutritional values

• filled: check if possible. You might want to correct the rounding issues.
• not filled: copy it from the image, if available.

Final check

Look at the product page and see if the ingredients are recognised. Edit if required. Report any unrecognised ingredients or try to work on the ingredients taxonomy.