Frequency aggregation
Frequency aggregation is the process of converting higher frequency data to lower frequency data. For example, a time series with a monthly observation frequency can be aggregated to annual frequency.
The method applied for gathering data for official statistics at the ECB are strictly defined in the manual and guidelines outlined in the statistics co-operation and standards. On the ECB Data Portal, you can do frequency aggregation transformations yourself in the various dashboards. For your selected time series, you can modify frequency aggregation based on already defined methods depending on the time series’ collection indicator.
- Average: mostly used for time series with indices.
- Sum: commonly used for calculation of flows.
- End-of-period: an accounting principle used when calculating stocks.
For example, the time series collection (identifier: CL_TIME_COLLECT) for the QSA dataset can be found under time series-level attributes of the data structure section (here). For a specific time series, such as Gross disposable income of households, you can find the time series collection by going to the time series page, clicking on the Menu button and selecting Show data information in the drop-down menu.
Show data information
Average
This method averages all observation values within the aggregated frequency. For example, for a time series aggregated from monthly to annual frequency, the 12 observation values of the monthly data are summed and then divided by 12.
Summation
This method sums all the observation points within the aggregated frequency. For example, for a time series aggregated from monthly to annual frequency, the 12 observation values are summed.
End-of-period
The last value in the period is the end-of-period observation value. For example, for a time series aggregated from monthly to annual frequency, the monthly value for December will be the annual value for that year.