For this first one, I think we could default it to repos with less than 95% completeness. Also, how hard would it be to make the repo_id's vertical, with a legend? Another solution that would maybe be easier, and of equal value, would be to show "number of repos above 99.5%, and number of repos below 99.5%. From there, maybe allow people to click on any repos in a paginated list? I don't want to make this a lot of work! The use case Is folks wanting to know PR data collection status, and high level/summary info by default (first), and detailed drill down for repos with lower completeness percentages (second) would, conceptually, be a better design. Maybe we can make this a software engineering project? ;)
These are good, and maybe address my thoughts above. I do think we will need to limit the n we display. Some of our installations are in the 20,000+ repository range:
Here's a case for using Z-Scores if I have ever seen one 🤣:
This would also require some kind of top n or other default, or user selection of repos when we get to any size. It could also be a pie chart that can be downloaded with a legend, possibly in JSON format. Just spitballing here.
=================================================
I am not sure what I am looking at here. That could be easily fixable with a legend. I think its saying, "we ran collection in February", but there's nothing new for grimoirelab-kibiter, for example, since ~october. I think if it was annoyed with divided by from repo_info, and paginated within the graph, that might be a good pattern of design to follow for issues and PRs, aside from/in addition to the top level summary of <total all prs/issues collected> divided by <total all pr/issue metadatacount>. Repeating myself, but that's the first piece of info I think people want.
For this first one, I think we could default it to repos with less than 95% completeness. Also, how hard would it be to make the repo_id's vertical, with a legend? Another solution that would maybe be easier, and of equal value, would be to show "number of repos above 99.5%, and number of repos below 99.5%. From there, maybe allow people to click on any repos in a paginated list? I don't want to make this a lot of work! The use case Is folks wanting to know PR data collection status, and high level/summary info by default (first), and detailed drill down for repos with lower completeness percentages (second) would, conceptually, be a better design. Maybe we can make this a software engineering project? ;)
These are good, and maybe address my thoughts above. I do think we will need to limit the
nwe display. Some of our installations are in the 20,000+ repository range:Here's a case for using Z-Scores if I have ever seen one 🤣:
This would also require some kind of
top nor other default, or user selection of repos when we get to any size. It could also be a pie chart that can be downloaded with a legend, possibly in JSON format. Just spitballing here.=================================================
I am not sure what I am looking at here. That could be easily fixable with a legend. I think its saying, "we ran collection in February", but there's nothing new for grimoirelab-kibiter, for example, since ~october. I think if it was annoyed with divided by from repo_info, and paginated within the graph, that might be a good pattern of design to follow for issues and PRs, aside from/in addition to the top level summary of <total all prs/issues collected> divided by <total all pr/issue metadatacount>. Repeating myself, but that's the first piece of info I think people want.