add aggregation type for temporal data

[sfluegel05]

Description of the issue

An aggregation type is usually part of the description of a time step. This appeared in #267, see #267 (comment) and #267 (comment)

Ideas of solution

We could implement this analogous to time stamp alignment:

• a class aggregation type that is a subclass of data descriptor
• individuals of this class that represent the different possible aggregation types

Aggregation types mentioned so far:

• sum/integrated
• mean
• instantaneous

Workflow checklist

• I discussed the issue with someone else than me before working on a solution
• I already read the latest version of the workflow for this repository
• The goal of this ontology is clear to me

I am aware that

• every entry in the ontology should have a definition
• classes should arise from concepts rather than from words

[carstenhoyerklick]

We should also add "maximum" and "minimum" to the aggregation types. A typical "maximum" are e.g. wind gust in a time interval. Similar could be maximum and minim prices in the time series.
A further option could be something like "number", e.g. if we count the number of flahes in a thunder storm within a time interval. Or maybe something how often a thereshold of a price level is surpassed.

[carstenhoyerklick]

And another one: "flags" could als be a possibility, e.g. for quality control flags of measurement time series.

Number could occur e.g. as the number of values used to create an average in an aggregation process.

[stap-m]

Also mean shold be called "arithmetic mean". Are other types of mean used for the aggregation, too?

Suggestion for a definition: An aggregation type is a data descriptor that contains information on the aggregation method applied on a data set.

We could add an example of usage here, too. E.g. temporal aggregation of continuous measurements for discrete a time series, aggregation of a data set with a spatial references

[stap-m]

Do individuals also get definitions @sfluegel05 ?

I'd say, "number" (e.g. of flashes) is the same as "sum/integral".

[sfluegel05]

So far, we have some individuals without definitions and some that have definitions. In general, I'd say definitions are not necessary for individuals, because there is just one of them.
In this case it might be nice to have some examples or definitions so everyone understands what is meant by a term.

[stap-m]

Defs for the instances. Apart from instantaneous, all other instances can be used for temporal AND spatial aggregation. Maybe we have to specify "spatial region" when classes for georeferencing are implemented. But for now:

• instantaneous: value measured at or modelled for exactly one 0-dim temporal region, referenced by a time stamp, e.g. to represent a time step in a time series
• integral: sum or integral of values, e.g. within a time step or a spatial region
• arithemetic mean: average value claculated by arithmetic mean over a set of data values, e.g. within a time step or spatial region
• minimum: minimum value within a set of data values, e.g. in a time step or spatial region
• maximum: maximum value within a set of data values, e.g. in a time step or spatial region

[stap-m]

For "flag" we should find another place than aggregation type, propably as own data descriotor subclass. Any ideas for a definition @carstenhoyerklick?

[stap-m]

I'd propose to implement what we have and discuss "flag" in another issue, so that we can close here.