Technical Tips are quick ways to improve your data which you can easily apply for your next publication. This time: what is the aggregation status in the indicator?

Aggregation status is a field which only has two options: yes (true) or no (false). If you set it to ‘yes’, you indicate that this indicator is measured in such a way that the data can be aggregated – added up. Over time, over regions, or others.

An example: say you work in Kenia and you have a result ‘increase girls school attendance’, with two indicators: ‘number of trainings given’ and ‘percentage of girls going to school’. The first indicator has three periods in which you count the number of trainings given in that year. Say in the year 2020 (year 1) you gave 13 trainings, in 2021 (year 2) you gave 12 more trainings, and in 2022 (year 3) you gave 10 more. Ideally, you record these separately for each period:

Project in Kenia

Period 1 – 1 january 2020 to 31 december 2020: actual 13

Period 2 – 1 january 2021 to 31 december 2021: actual 12

Period 3 – 1 january 2022 to 31 december 2022: actual 10

Aggregation status is ‘yes’ in this case. In dashboards, the ministry and others can aggregate these data to a total of 35. But there are more possibilities. Perhaps you are part of an alliance and your partner has done these trainings in a different country. You worked in Kenia but your partner worked in Zambia. Their indicator periods are:

Project in Zambia

Period 1 – 1 january 2020 to 31 december 2020: actual 5

Period 2 – 1 january 2021 to 31 december 2021: actual 7

Period 3 – 1 january 2022 to 31 december 2022: actual 4

Clearly they worked with fewer trainers or a smaller area with fewer schools. They set their aggregation status to ‘yes’ as well. Now the ministry can use these data to conclude: there were a total of 16 trainings in Zambia.

But we can also see: there were 19 trainings in the year 2021: 12 from you in Kenia and 7 from your partner in Zambia. You can see how useful it is to be able to set the aggregation status to ‘yes’, if it applies.

It doesn’t always apply. Let’s look at the other indicator: ‘percentage of girls going to school’. This has the same periods and has the following data:

Period 1 – 1 january 2020 to 31 december 2020: actual 40%

Period 2 – 1 january 2021 to 31 december 2021: actual 50%

Period 3 – 1 january 2022 to 31 december 2022: actual 65%

Clearly, you can’t add these up and say that over the three years, 155% of the girls went to school. Aggregation status is ‘no’ here. In this case, that’s because it’s a percentage. Percentages can almost never be aggregated, so a rule of thumb is here that you should set the aggregation status of percentages to ‘no’.

The Ministry of Foreign Affairs in the Netherlands considers it extremely useful to be able to aggregate data wherever we can, so make sure you tell us for which indicators this is possible by setting aggregation status to ‘yes’ for those indicators. If the data is not suitable for aggregation, set it to ‘no’. Even better is to also record the data in such a way that the indicator becomes suitable for aggregation. In the first indicator, you could have worded in such a way: ‘total of trainings given in this region since the start of the project’. You would then have recorded the periods as 13 for year 1, 25 for year 2, and 35 for year 3. This is technically not incorrect, but not desirable because the data cannot be used for all purposes we would like to use it for. So if you can, word it so it is suitable for aggregation, and if so, tell us by setting the aggregation status of the indicator to ‘yes’.

Check out other technical tips from the Helpdesk Open Data! And if you haven’t yet, subscribe to our newsletter for future quick fixes which will improve your data instantly.