Sunday, June 04, 2023

Covid enquette in the Netherlands ..


Dutch politicians will have their day in an "enquette commissie" coming up with their version of what there is to learn from the COVID-19 pandemic. The result will probably mirror the composition of the mainly right wing commission. It is not as if we do not know what there was to learn; a recent research paper commissioned and paid for by the Dutch government lays out the effects of vaccination on mortality in the Netherlands. It does even refer to a study on religion and vaccination coverage in the Netherlands.

Given the composition of the committee, they will have a hard time convincing people of their outcome. Nobody is really interested in my opinion, but I do believe in proper science. Given the qualification of the people writing the commissioned paper, I added the paper to Wikidata, author strings were replaced by author identifiers, known publications were linked to the authors, I linked the cited papers and ensured that the primary author is known to Wikidata as well. The result is that anyone can find this enriched information in its Scholia.

Given that some Dutch politicians express that they not understand the scholarly process, the best I can do is open up information that has a scholarly foundation, particularly in a field where I do trust politicians to come up with a report that reflects their political bias.



Thursday, May 18, 2023

For Dr @ashadevos there are 14 @Wikipedia articles


Hardly a "woman in red", Dr De Vos has many accomplishments chronicled in these Wikipedia articles. She presents herself with her colleagues on Facebook and, the graph of her co-authors should paint a similar picture, initially it did not. At first there were only a few publications to her name, they have been expanded to 26 at present. It introduced many co-authors and there are now some 112 co-authors missing.

Obviously, there is much more that could be done. Adding more papers and co-authors adds complexity to the Scholia of Dr de Vos. More distinctions could be added, talks at conferences and papers that were cited. I typically restrict myself to papers with a DOI and authors with an ORCiD identifier as they have the biggest network effect. 

I was reminded by Greenpeace that some people give themselves nothing for their birthday. So I updated this Wikidata item. Who will notice or care.. Like Greenpeace, Dr De Vos cares about whales; it is her specialty.


Sunday, May 14, 2023

Gender balance at @Wikipedia, deletion; a rear guard action.

 A recent Wikipedia Research article aims to prove that the English Wikipedia deletion process is not biased. For some that is a loaded question because it  centers on the question if Wikipedia is equitable.

As so often the article is all about English Wikipedia and it has its own bias. English Wikipedia does not serve half the public of the Wikimedia Foundation and much of the other half does not read English. The gender balance in English Wikipedia is however improving; the percentage of articles about women is slowly but surely increasing.

At issue in the article is whether the English Wikipedia deletion policies effectively harm gender and race biases. Obviously there are more biases; you may be male and white but when you are not from an Anglo-american background chances for Wikipedia recognition are slim. When you care to research this, check out Wikidata, it includes a super set of what Wikipedia includes and it is biased in this way as well.

When a Wikipedia article about a scientist is deleted, it does not follow that its Wikidata item is deleted and given enough identifiers, it is likely that its related subset increases over time tilting the "notability" balance. Even so, many important scientists are "scientists in red", an example is Prof Emily Fairfax her prominence is for instance in her explaining and demonstrating that beavers feature prominently in the fight against forest fires

When English Wikipedia defends its own policies, it follows that they rely on the base assumptions in those policies. When those assumption are questioned, their arguments are lost. Given that English Wikipedia represents a subset of "the sum of all knowledge" that is included in Wikidata, it follows that much of Wikipedia can be understood from such a perspective. 

Wikidata has no "red links"; when a relation exists for an recipient of an award, there must be an item for both the award and the recipient. Wikipedia has one link in black to the "SIRS Lifetime Achievement Award". while Wikidata has a link to all recipients. They are linked to identified publications and other awards and consequently the Scholia for the award is really informative. 

Based on information like this improved information is available that must wait for a Wikipedia volunteer. English Wikipedia is a victim of its success, it cannot fully maintain its information. The same can be said for Wikidata. It is however a superset and it does not necessarily require a mastery of English.

With new technologies becoming more relevant, there is an avenue to improve the quality of any Wikipedia, inform people based on the data in Wikidata and improve on the quality of the information that we provide. 



Sunday, April 23, 2023

Analysis of a Wikipedia article

For professor Gerhard Gries there is a Wikipedia article and as a consequence there was a Wikidata item. Actually there were two; they have been merged. Many articles have been attributed to professor Gries and a Scholia template was added to the article.

Professor Gries has his own lab: the "Gerhard and Regine Gries Lab". There is no article for his wife and, there was no red link. There was no Wikidata item for his wife but there were 59 links indicating her as a missing co-author. She has now 68 articles linked to her as an author or co-author. Regine Gries is now a "woman in red".

The article states in two places that he is a "Fellow of the Entomological Society of America" (2019). It is in the text and it is in a Wikipedia category. Relevant are two additional lists; the website of the Entomological Society of America and the result of a query at Wikidata (shown in a Scholia). All these lists are incomplete, the fellows of 2022 are not yet included on the SoA website; they can be found in a different place. The Scholia has been added to the category article on the English Wikipedia; in effect you will find several "fellows in red". 

Three additional awards are listed, awards that have no category on the article. One is for the "Fellow of the Entomological Society of Canada", its reference may be found here. Another one is for the Nan-Yao Su Award and finally the Gold Medal of the Entomological Society of Canada had to be added to Wikidata. There are always more "award winners in red" to be found or to be added.

Every list, every category in Wikipedia is likely to be incomplete. What we do not know may be relevant. With data on missing articles available in Wikidata, there are more more options to make Wikipedia more inclusive.


Saturday, April 22, 2023

he, she, they/them

One ambition at Wikipedia is to have more articles about women.  The Women in Red project does really well, slowly but surely the balance between articles about males and females is improving. How do we know this: in Wikidata we have a database we can query and it shows over time.

Obviously, both many deserving men and women could get an article in future and particularly many scientists are already known in Wikidata through their publications. So how do we know the gender of these scientists? Because of a name like Emma or Janice it is likely a woman.. Not a precise method particularly for those people who identify themselves in a different way. Google scholar or Twitter often shows a picture and that is not fool proof either. 

The dilemma is in two ways: manual entries are open to errors in the first place. A six percent error rate is to be expected in any edit and anyone is kindly requested to fix what should be improved; Wikidata is rich in alternatives for male/female identifiers. The alternative is that we do not add a likely gender. This results in no awareness of the composition of the co-authors of an author. No awareness of the volume and balance of people who do not have an article yet.

I think that a male author with only male co-authors is problematic in and of itself. Quite often it is just that no attention was given to female co-authors so I often remedy this by giving attention to them. I add them to Wikidata, look for an ORCiD identifier, a Twitter handle a Google scholar profile. The effect is not only apparent for the male author, but it has an effect on all the co-authors for the newly registered author.

The issue I have is, I see no solution for the dilemma of a gender balance in Wikidata. What I do know is that Wikidata is a collaborative project and anyone is kindly requested to make it as good as it can be.

Thanks, GerardM

Saturday, April 08, 2023

@WDScholia, proficiat Prof @DrSarahNull

 On Twitter it was announced that Sarah E. Null will be a professor at USU Watershed Sciences. Hydrology will prove to be a subject of even more importance. This subject has my interest so I checked out her entry in Wikidata.

Identifiers are really important, so I added her Twitter handle and her Google Scholar ID. Her Scholia showed what papers were known and I clicked "improve data" to disambiguate additional publications and link them to her Wikidata entry.

Google Scholar provides information of her publications and I added one paper to Wikidata; "Prolonged and Severe Drought in the Most Dammed Tributaries of the Lower Mekong Basin". It has 8 authors and only four have an ORCiD identifier, three did not have a Wikidata entry. Now they do and their ORCiD identifier makes them unique.

Typically I add several publications when I concentrate on a person. When I stop, I often revisit the "improve data" part of a Scholia to add missing co-authors. Of particular interest are the ones who are missing the most. They have the biggest impact on the "Co-author graph", making it more complex. Jay R. Lund is missing as a co-author ans so is J. R. Lund. Wikidata knew a "Jay Lund" and a "Jay R. Lund", they have been merged and I was then possible to attribute the missing publications for Jay R. Lund. After all this, Prof Null is no longer the only known author on seven papers. :) A bit diapointing so I added one other co-author for now; Marcelo A. Olivares. Others will creep in as more work is done in the future.