Aggregation and disaggregation
Goal: Bring data into weight-based units (tonnes) for the system boundary and the same reference year.
Situation: The data does not exist at the right spatial level, the needed level of detail or the right time frame.
You can aggregate or disaggregate other data. Aggregation and disaggregation, as the names suggest, is the grouping or splitting up of data, respectively.
Aggregation needs to occur if data exists on a more detailed level that is required or useful. Examples:
- the spatial data is on a neighbourhood level, then it has to be added up from all neighbourhoods to the city.
- trade data exists on a too detailed temporal scale such as for quarterly or monthly time frames, then it should be aggregated to a yearly time frame.
- quantities of produce are provided in greater detail than necessary, for example that it is distinct as far as carrots, while an aggregated level such as vegetables is required.
In these cases, a simple adding up of smaller groups is necessary to achieve the goal.
Disaggregation has to be carried out if the information is too general, for example
- there is data on a material group that combines materials in a different way from the EW-MFA materials.
- the data is given for a range of five years, instead of single years.
In these cases, informed assumptions have to be made, for example that the material group is made up of a certain percentage of the material of interest (e.g. 20% of imported food are cereals).
Outline of the video
- 0:07 What do we mean by that? Whenever you carry out a material flow analysis or circularity assessments, you generally get information data. But you don't always get it at the right spatial level or detail level that you require, and sometimes you need to aggregate it = put things together in order to get to the right format that you need or disaggregate it.
- 0:31, So take something from the big and then split it into many smaller pieces in order to get to the level of granularity that you need.
- 0:43, What we need: We are here in the middle and that's what we want. We want a certain type of aggregation or disaggregation. Our role is to put stuff together or take things apart in order to get the information that we need at the level of granularity that we need.
- 1:12, examples
- different types of households in Morocco: So that person went and made some typologies of households and then measured the flows per typologies of households. He summed up the number of households per type of household in order to get to the level of the city. → 1:59 If you cannot have information for your own city, you can make some typologies and do some surveys or find information for a smaller spatial scale and then add them up to get to the bigger scale.
- 2:13, other example of aggregation can be temporal aggregation. We are aiming for a year, a reference year, and we want to have tonnes or kilotons or kilograms for one specific year. You get information at the monthly, daily or weekly, basis or frequency. You need to sum them up in order to arrive at a year. Example of monthly water caught in a place in France.
- 3:11, examples of disaggregation, on the left side you have a screenshot about how much materials are exported from Brussels. We have the total amount of materials that are exported per country, we don't know what is exported. → I went and looked at the national level. You take the level, you take the share of the composition of materials exports from your country and then you apply to the total amount that you already have, which this time is accurate at your level.
- 5:25, One other way to do it is for instance, we had the same issue for different types of transportation. So had accurate data and disaggregated data for let's say road transportation, but not for rail transportation.