feat: addition of accel data

[m-muecke]

Closes: #1

[fabian-s]

covar <- readr::read_csv(here::here("data-raw", "covariate.csv"), )
activity <- readr::read_csv(here::here("data-raw", "activity.csv"))

can you add those files to data-raw please?

[m-muecke]

covar <- readr::read_csv(here::here("data-raw", "covariate.csv"), )
activity <- readr::read_csv(here::here("data-raw", "activity.csv"))

can you add those files to data-raw please?

covariate.csv is not a problem since its only 2 kb, but activity.csv is 230 mb

[m-muecke]

covar <- readr::read_csv(here::here("data-raw", "covariate.csv"), )
activity <- readr::read_csv(here::here("data-raw", "activity.csv"))

can you add those files to data-raw please?

covariate.csv is not a problem since its only 2 kb, but activity.csv is 230 mb

@fabian-s whats your preference, since the activity.csv file is too large?

[fabian-s]

sorry, need to think about this a little more --
seems like a general question on how/where/whether to preserve the raw, un-pre-processed data for a data package like this.

one idea would be putting the resources that belong into /data-raw into a zenodo or figshare deposit and accessing it from there in a reproducible manner.
other idea is using Git LFS, but i just know that that exists not how to use it.... 🙈

[m-muecke]

sorry, need to think about this a little more -- seems like a general question on how/where/whether to preserve the raw, un-pre-processed data for a data package like this.

one idea would be putting the resources that belong into /data-raw into a zenodo or figshare deposit and accessing it from there in a reproducible manner. other idea is using Git LFS, but i just know that that exists not how to use it.... 🙈

Makes sense, there is also the GitHub large file storage: https://docs.github.com/en/repositories/working-with-files/managing-large-files/about-git-large-file-storage

[m-muecke]

I think we should move forward with using GitHub LSF, as it looks to be very well integrated with the Git repo.
Another option would be to store the files as parquet. The activity CSV shrinks from 200mb to 34mb using Parquet, and I don't think there is much, if any, benefit in keeping the actual CSV.

Having a tidyfun zenodo would actually also be quite lovely. Downloading the records from R is also very straightforward, for the Monash time series zendo I've written a script here