Every time series has a "native frequency" and Data Buffet is able to "convert" among frequencies. Which frequencies are available, and what does a "subfrequency" setting like "Annual (April)" or "Weekly (Sunday)" mean?
Frequency and subfrequency
A time series exhibits these properties:
- Its measurements are made at regular intervals, i.e., the series has a fixed periodicity or frequency.
- It can be converted from its native frequency to other frequencies.
To perform task (2) you need to decide how the native frequency is aligned to the output frequency. For example, 12 monthly periods are aligned to a year such that "month 1 (January)" is at the start, and "month 12 (December)" is at the end -- but that's just a standard calendar year. Conversely, a fiscal year might run from April to March, i.e., its phase is shifted by three periods. This is the role of subfrequencies in Data Buffet.
In Data Buffet, a subfrequency is named by the period in which it ends. For example, "Annual" is shorthand for the default "Annual (December)", whereas "Annual (March)" describes the above fiscal year.
Because we collect data internationally, from sources that use different fiscal years or collection phasing, Data Buffet stores data using subfrequencies. To convert variegated time series to a compatible frequency, our modules offer the most common subfrequencies as options.
Frequency options
In all Data Buffet modules there is a drop-down control labeled "frequency" with these options, in alphabetic order. "Default" means "use the native frequency of the series", which is a useful distinction if you individualize the series in a basket.
- Default
- Annual
- Annual (April)
- ... 10 others ...
- Bi-Monthly
- Bi-Monthly (November)
- Business Daily
- Daily
- Monthly
- Quarterly
- Quarterly (November)
- Quarterly (October)
- Semi-Annual
- Semi-Annual (August)
- ... 4 others ...
- Weekly (Friday)
- ... 5 others ...
- Weekly (Wednesday)
By order of decreasing frequency (or increasing period), this is what they mean:
| Subfrequency | Period count | Phase pattern or other comment | Concept character |
|---|
| Daily |
7 / week |
Used if continuous data, or for non-European work weeks. |
D |
| Business Daily |
5 / week |
Used for Monday to Friday. |
D |
| Weekly (Friday) |
52 / year |
Week ends on Friday. |
W |
| Monthly |
12 / year |
|
M |
| Bi-Monthly |
6 / year |
Jan-Feb, ..., Nov-Dec |
M |
| Bi-Monthly (November) |
6 / year |
Dec-Jan, ..., Oct-Nov |
M |
| Quarterly |
4 / year |
Jan-Mar, Apr-Jun, Jul-Sep, Oct-Dec |
Q |
| Quarterly (November) |
4 / year |
Dec-Feb, Mar-May, Jun-Aug, Sep-Nov |
Q |
| Semi-Annual |
2 / year |
Jan-Jun, Jul-Dec |
S |
| Semi-Annual (August) |
2 / year |
Dec-May, Jun-Nov |
S |
| Annual |
1 / year |
Year ends in December. Standard calendar year (CY). |
A |
| Annual (April) |
1 / year |
Year ends in April. |
A |
A small number of Data Buffet series are stored with native frequencies other than those listed above, which will display in Mnemonic 411 and basket output, but we don't provide the option to convert to those frequencies. For example, "Biweekly A".
Data with a "censal" frequency is collected once every several years (five, ten, possibly irregular). We store this in an annual time series with "ND" values for intercensal years, and name the series with a "C" frequency character on its concept code.
Basket output
Since you can create a basket that mixes different native frequencies, you have these options:
- Each contiguous group of vectors (rows or columns) that shares a native frequency will be labeled with a date vector (default behavior).
- Show a date vector alongside every series.
- Convert the entire basket to a common frequency.
Location in each module
- View: Frequency
- Basket editor, grid mode: Frequency column » drop-down control
- Basket editor, power mode: Frequency
- Chart editor: Options » Output » Data » Frequency
- Map editor: Options » Series » Options » Frequency
Examples
Demographic and economic activity is continuous, and statistics necessarily choose to represent the activity at a particular moment (a "stock" measure, such as population, inventory or bank balance) or over an interval (a "flow" measure, such as births or transactions). Sometimes there is a natural best choice, but some indicators have equally valid alternatives (monthly value for daily stock price: average over the month, or end of the month). To learn which, check the "observed" property of a series and consult the dataset background (both in Mnemonic 411).
For example, the subfrequency "Weekly (Thursday)" may indicate:
- Activity is measured on Thursday
- Sum of activity over the Friday-to-Thursday interval
- Disseminated on Thursday but represents a rolling window longer than a week
See also
Using Data Buffet:
Example datasets that use subfrequencies:
Updates
- Aug 2017 - Initial version
- Jun 2019 - Additional subfrequencies, explanatory table, examples thereof