Make metrics above installation map on dataverse.org not misleading

[jggautier]

The count of dataverses, datasets and file downloads above the Dataverse installation map look like they're from all installations, but they're from Harvard Dataverse only. We have no automated way to update those metrics with counts from the known installations.

How can we make those numbers not misleading?

(Sorry for the weirdly-phrased titled. Trying not to use it to prescribe a solution.)

[jggautier]

Can we remove those numbers for now (and just have the count of installations in the center)?

[pdurbin]

Can we remove those numbers (and just have the count of installations in the center)?

@jggautier I love this idea.

[pdurbin]

The numbers are misleading. Check out this tweet from Jochen Apel at https://twitter.com/Duhem_/status/935133423877976065

"Quick, why does this website https://dataverse.org/ say that there are~50k datasets over all installations if Harvard Dataverse alone lists ~75k datasets?"

[pdurbin]

See this post for some confusion over the gray dots: https://groups.google.com/d/msg/dataverse-community/eQlSLFgzQXI/t6FANVynBgAJ

I think we should simply remove the gray dots for now. Like @jggautier said above, let's just show the orange dots and list the number of installations. The three non-installation metrics (dataverses, datasets, and files) are causing confusion. Perhaps in the future, we can try again with metrics that are not so specific to Harvard Dataverse, but rather reflect the entire Dataverse community. On my "Flagship Bias" doc at https://docs.google.com/document/d/1-LtDv_Yaf_3EPuMWcg5ZxH7M2ACrmCzPzgmbwbt8gMQ/edit?usp=sharing this is what I wrote:

Metrics also appear above the map on the project home page. The number of installations around the world is fantastic to highlight but the number of dataverse, datasets, and downloads apply only to the "flagship" installation.

[pdurbin]

I just approved pull request #66. The change makes sense. @mallove simply commented out the inaccurate metrics. Thanks!

[pdurbin]

This looks so much better:

Thank you, @mallove !!

@jggautier you opened this issue. Are you ready to close it?

[jggautier]

Closing! Thanks @mallove!

[pdurbin]

It was noticed during SpinachCon 2018 ( IQSS/dataverse#4505 ) that the misleading numbers are back so I'm re-opening this issue. Here's a screenshot from today:

[pdurbin]

I spoke with @djbrooke today about this and he said I could go ahead and create an issue in the main issue tracker so I just did: IQSS/dataverse#5429

Closing this one in favor of the new one.