search

WDscholia / scholia

Wikidata-based scholarly profiles
https://scholia.toolforge.org
Other
226 stars 81 forks source link

SPARQL for venue_co-author-graph - could line 18 limit be extended to 150 #1768

Open HelsKRW opened 2 years ago

HelsKRW commented 2 years ago

The SPARQL for venue co author graph at https://github.com/WDscholia/scholia/blob/master/scholia/app/templates/venue_co-author-graph.sparql has a limit of 15 on line 18. I am finding that this means some author collaborations are not displayed. If line 18 is changed to 150 then additional collaborations show.

To Reproduce Steps to reproduce the behavior:

  1. Go to https://scholia.toolforge.org/venue/Q97011661, where author collaborations above 4 authors are not displaying.
  2. see https://w.wiki/4iKs where the limit appears on line 19 of the SPARQL, and if changed to 150 results can then accommodate an article with 9 authors as shown in the results.

Expected behavior I hoped the venue co author graph would display collaborations for multiple authors and wondered if it was possible to change this in the SPARQL being used on Scholia so that a greater number of collaborations will display on the Scholia page.

Desktop (please complete the following information): I am working on a laptop with Chrome Browser

Additional context This is my first time using GitHub - I hope this has been an appropriate way to communicate. The way in which Scholia is visualising data pulled from Wikidata, and particularly the co-author collaboration graph is amazing! If it's possible to increase the number of author collaborations it's possible to visualise this would be really helpful. Thank you!

Daniel-Mietchen commented 2 years ago

Thanks, @HelsKRW , for the suggestion. Yes, this is a very appropriate way to communicate with us. I have relabeled your suggestion as an enhancement, though, not a bug, because the current LIMIT 15 is the result of comprehensive testing. For instance, if you replace LSE Public Policy Review (Q97011661) with Physical Review B (Q2284414) (where collaborations with hundreds or even thousands of co-authors are common), then your query variant with LIMIT 150 times out, as per the screenshot below: Screenshot 2022-01-24 at 19-53-49 Wikidata Query Service

With the current LIMIT 15, the entire profile for Physical Review B renders just fine: Screenshot 2022-01-24 at 20-02-36 Physical Review B - Scholia

So what we could do perhaps, though, is to encourage users more to explore the query on their own, perhaps by changing the Wikidata Query Service label to something more along the lines of Explore this query via the Wikidata Query Service.

HelsKRW commented 2 years ago

Thanks @Daniel-Mietchen for your kind response and explaining this so clearly. I now understand the 15 limit, and like the idea of encouraging users to explore further on their own. I'm quite new to Wikidata, and very new to Scholia and SPARQL so it took me a while to work out that changing the limit was the way to change the results (initially I tried adding author3 author4 etc at various points in the SPARQL). If other users are also likely to be novices it may be useful to provide some brief guidance on how they could explore further - but it may be that most users already have more advanced technical knowledge. Thank you again for your help - I really appreciate it.

fnielsen commented 2 years ago

Thanks for you report @HelsKRW. The Scholia queries are sometimes not optimal. We always have to keep in mind that we are querying over many authors. Also we would like to have the plots more interactive, so you, e.g., could change the limit of the number of coauthors shown.

HelsKRW commented 2 years ago

Thanks for your message @fnielsen - I like the idea of making the plots interactive. Scholia is a great resource! Thank you.