Create generic data type(s) for internal usage

[enriched]

Concept is to identify commonalities in the data so that we can generically combine data from different sources. To do this we need to identify the key differences and similarities in the data that we are consuming from external apis as well as have a roadmap for what we intend to consume.

[TheTedAdams]

I think the important thing is that every metric is keyed by date. Then if metrics are mostly numeric, we can always look for correlations over the date axis

[enriched]

Off the top of my head:

• timeseries
• time points
• histograms
• any n-dimensional matricies?
• scalar int
• scalar double

[enriched]

fitbit units (per the api): weight measurements liquids height elevation duration distance blood glucose

[enriched]

Google fit is probably a well thought out example for exercise data

[enriched]

Using a dimension, metrics, and segment api could allow us to have a dynamic system of available metrics for a given data stream. We could query that stream to find the available metrics and create views accordingly.

[enriched]

https://developers.google.com/analytics/devguides/reporting/core/dimsmets

[enriched]

I think that the best pattern might be a entity component system. That allows for discovery of what is available as you traverse the tree of datasets.