Comparing apples and oranges of data

Person holding apple fruit - Andrea P Coan

According to the common idiom, you cannot compare apples to oranges. The idiom exists in other languages as well, although in French it’s about apples and pears – comparer des pommes et des poires. This must have something to do with calvados making process. Both of them can be used for calvados, but the fruit varieties and their ratios are heavily controlled by AOC, the French certificate for traditional agricultural products from certain areas.

The idiom refers to the apparent differences between items that are popularly thought to be incomparable. In real life, whether comparing different things makes sense or not, is always defined in the context. If your plan is to prepare some sweet tasty nutritious juice, both of them are good. When you’re making a nice fruit cocktail both them might be good additions. However, if you’re baking an apple pie there are quite strict rules. You need to use acidic apples, and pears (and definitely not oranges) are not an option.

The idiom relates to the main goal of Platform of Trust ontology, which is data harmonization. The harmonization can be divided into four parts that are shown in the picture.

Data models help harmonization of data

An obvious benefit of harmonization is to enable interoperability between different data sources. When real-world things are classified with a common data model, they become semantically interoperable. In practice, this means you can actually compare apples with oranges due to the model classifies both them as edible fruits. And when sometimes you need to pick only the apples from ten different fruit baskets, it is easy because an apple is classified exactly the same way in all the baskets.

“An obvious benefit of harmonization is to enable interoperability between different data sources.”
Panu Pitkänen
 – Ontology Lead, Platform of Trust

You can use the same picking technique for each of the baskets, or even better, pick all the apples from ten baskets at once. Interoperability is a benefit on the general level, as well as individual level. Being able to identify certain individual, for example, a person with a national identification number, make it possible to combine data of this person across various data sources.

Data values are measurements of devices, physical quantities, or other subjects of interest the data contains. Harmonizing them means they are converted in the same format and using the same measures of unit. In practice, this is for example that UK and US customary units are converted into the metric system – and vice versa, depending on the customer’s requirements.

That is data harmonization in a nutshell.

Take a look at our ontology through a convenient tool at

Panu Pitkänen 150 x 150.png

Thank you for reading! Please, comment below.

I'm Panu Pitkänen and I’m the Ontology Lead in Platform of Trust. In June 2020 we launched a new version of our ontology that is the information model (or data model) used for harmonizing data. The new version has several improvements, and I will tell about them and other ontology related topics in this blog in a very un-technical way. Hope you enjoy them!

Feel free to contact me at +358 400 665 347, or in Twitter @panupit