Journal/Pine nuts
A journal has been kept of the cleanup process of this category in order to get a better feel what needs to be done, issues that arise and to assess the impact. The following steps were followed:
Definition
The first step is to assess what we are talking about. The category Pine Nuts has 297 products on 15 nov 2024. It has a single subcategory Shelled pine nuts, which has 50 products. I did not know that these were sold unshelled?
From a botanical side these are also called seeds, but then any nut is a kind of seed. There seem to be many variants of these edible seeds from different trees. In fact there is category cedar nuts, which should be put under this category.
Scrolling through the products in the category, all products seem similar. There are no obvious other related categories. Thus anything we recognize as a pine nut we can assign to this category, through the Hunger game, which is the next step.
Hunger game
Hunger game has some 100 images of supposedly pine nut products. Some observations: can be toasted, from Ceders, Spanish?, Siberia? (Pinus Siberica), China. Images from the nutritional table might be correct, but I can not confirm those.
At the end we have 555 products in the category. maybe we need to do this again tomorrow.
In noticed some words that might indicate an origin or a specific category. It might be nice if we have alle the text extracted from image, so that we can search for those words.
The next day (16 nov 2024) I redid the search, but no additional products were to be done.
Categories
There are 46 categories used for describing pine nuts. Some of these are not relevant and the products need to be edited. A lot of US products have been put under Snacks: removed.
In total 55 products have been edited. Now 552 Pine nut products.
Ingredients
There are 207 ingredients for the pine nuts category. We expect only a few, so a lot to go through. It seems we need to add the tree species as category: Pinus Sibirica, European Pine and Pinus Koraiensis.
In Spain they sell sugar coated pine nuts, moved to Nuts and their products.
After editing 58 products, we are left with 17 ingredients. Most of these have been added as new synonyms. There are only ingredients for some 170 products.
There are 68 products where the ingredient image has been selected, but not the data extracted. Let's have a look.
70 products edited, completed and/or repaired. At the same time I could add origins and improve categoristaion.
Languages
Next step is to make sure that the ingredients and product names have been set to the correct language. The easiest is to download the data as csv and go through the product names columns. I do only English, Dutch, French, Spanish, Italian and Portuguese. Possible errors are: the main language is set wrong, a language has been added that does not exist, or just the wrong language added. While editing I can add origin or adapt category.
This required editing 94 products! Discovered more toasted pine nuts, so that is really a category. But this means that an Unroasted category is required.
Needed to repair some 10 products for which I entered the wrong new category name (Pinus Sibirica).
Quality
Next step is moving to the nutritional values. The first indication is given by the quality errors.
Thus 9 products could be repaired, 3 not.
Graphing
In order find other nutritional error we need to create graphs of each nutrition versus the energy and inspect the outliers.
Quite a few US products which are totally off: deleted.
Some 5 products have to much salt for the category. It could be saled pine nuts, but nor the ingredients nor the product name indicate this.
56 products needed correction.
It seems there are two subgroups, but need to investigate further.
Wrapping up
Now that we have a good impression on the ingredients and subcategories, I can add those to the taxonomy:
- the expected ingredient for the whole category and subcategories is pine-nut. No other ingredients have been detected.
- We can add Korean, European and Siberian pine as subcategories, which are mutually exclusive (oppposites).
- All the pine nuts are shelled, so we could remove this category (but left at the moment);
- Added roasted pine nuts and thus also Plain pine nuts (unroasted)
Should one nuts be considered as Nuts for the Nutriscore calculation?