Is there a Julia package that can decompose a time series data into trend, seasonality and random?

Hey all,

I wrapped the x12 R package in https://github.com/julianjohs/X13! It is still in a very early stage of development and I’m working on a proper documentation. Anyway, as of now, X13 provides the basic features of the wrapped R package, i.e. the X13-ARIMA-SEATS decomposition/seasonal adjustment. I hope people reading this thread later will find this useful!

7 Likes

As a big side note, I interviewed/gave a talk on Julia at the Census Bureau a few years ago and they were entertaining the idea of porting X13 to Julia. I think the situation was that X13 is written mostly by one guy who was about to retire and they wanted to make sure they could hire people to maintain it.

11 Likes

I was also looking for one way to decompose time series in Julia so I implemented STL here:

The package also has a plot recipe that mimics the R output when plotting STL objects, however there are some differences to make it closer to the choices made in the original paper for the representation.

co2.

Merry Xmas!

14 Likes

StateSpaceModels.jl is a really nice option too.

6 Likes

I was thinking of creating an artifact with the original FORTRAN code, does anybody know the steps to do that?

Maybe take a look at Home · BinaryBuilder.jl Also here Calling C and Fortran Code · The Julia Language for some ideas on how to do it.

1 Like