More representative example dataset

[peterdesmet]

@PatrickAJansen suggested to use a real example dataset in camtraptor, rather than current artificial one from Camtrap DP. The latter was reduced and tweaked to cover as many use cases as possible. While that one is fine for camtrapdp, I agree that camtraptor could benefit from a real one.

Which dataset?

Patrick suggested to use a Panama dataset included in activity, but that one is 1) not available in Camtrap DP format and 2) a living dataset. Better would be to use a real published dataset.

https://album.wildlabs.net/dataset/search gives an overview of dataset, but I don't think any currently meet the criteria of 1) covering multiple sites, 2) including working links to images and 3) being relatively small.

@jimcasaer is there any small dataset we could publish?

Impact

Switching the example dataset would obviously have an impact on tests, examples and vignettes.

[damianooldoni]

Indeed, switching to a better example dataset will have a big impact on a lot of documentation and tests.

A solution we could implement is to leave the standard Camtrap DP example dataset for functions documentation and tests, while using a real example dataset for the vignettes.

[jimcasaer]

I would suggest to take the GMU8 dataset that can be found on GBIF and to make a selection for several months out of this dataset as example dataset. Maybe also make a subset in the number of locations but at least be sure to include all used locations twice in order to test certain features of camtraptor.

[jimcasaer]

Zonien could also be used given that there the CT are not moved every month - but this dataset is not yet published