Everything is a time-series. Where are the corresponding abstractions? How to … - Julien Perro

Julien Perro introduces a Scala library designed to abstract and combine heterogeneous, misaligned time-series data sources as functions of time.

Everything is a time-series. Where are the corresponding abstractions? How to (almost) hide away a heterogeneous reality. - Julien Perro Did you ever have to combine multiple sources of time-series like data that were: - misaligned - had different sampling frequencies - a lot of missing measures? This kind of situations generated enough frustration for us to come up with an abstraction to hide all this under the rug: We present a small library, [scala-timeseries-lib](github.com/Sqooba/scala-timeseries-lib), that abstracts the discreet and flaky nature of time-series data away, looking instead at them as functions of time, which can be arbitrarily combined. Multiplying, summing, ... over your heterogeneous series is a breeze, while you can build series of any objects and define your own operators between them. In essence, you can do: `val externalTemp: TimeSeries[Double] = ...` // load data from your favourite weather service `val internalTemp: TimeSeries[Double] = ...` // load data from your indoor sensor `val difference = internalTemp - externalTemp` // And very easily compute the difference. ... and stop worrying about the details once and for all.