Update the wikipedia pages of scikit-learn

[francoisgoupil]

We could probably improve the scikit-learn wikipedia pages in terms of update and structure/content.

Why?

• I reckon it is a very important place for people don't knowing much about scikit-learn and I am sure we can do better
• I believe we could also better emphasize on the packaging and maintenance efforts.
• We could talk more about the community and its governance, give some key figure...

Action: Having in mind the different persona of readers, I think we could also come up with a new plan and structure with high to low level of information granularity.

[ogrisel]

I believe we could also better emphasize on the packaging and maintenance efforts.

I am worried Wikipedia is the place for this and we should refrain ourselves from using Wikipedia to promote scikit-learn but keep a very neutral, informative tone. The Wikipedia page of scikit-learn does not belong to the scikit-learn project.

[ogrisel]

But it's true that the contents of this page deserves an update and corrections:

French Institute for Research in Computer Science and Automation in Rocquencourt, France

For instance, the Inria Parietal team started in Saclay and was never based in Rocquencourt as far as I know.

[GaelVaroquaux]

But it's true that the contents of this page deserves an update

Also, it's currently very short with little on it. This gives the impression of a project that is minor, not very active, and not very used. It's important that we convey more dynamism (actor such as Microsoft use this impression on scikit-learn to convey the message to decision makers that it is a dead projects, and that they should focus on AzureML, I've seen this happening just yesterday in a private meeting)

[reshamas]

Wikipedia is legit and an open source project. The number of users is staggering. It is considered a credible source by many users, including Funders!

Also, it's super important to keep in mind that people who make funding decisions or doing research on potential projects to fund are not often users of the tool, and so even if scikit-learn is the number 1 python machine learning library, not everyone knows that. (Another reason why open source projects suffer from lack of funding.)

scikit-learn doesn't need to own the page to utilize and benefit from it. It's referred to as "earned media", which is quite beneficial.

scikit-learn is 10+ years old and our blog is < 1 year old. An important question to ask is how would everyone (or maximum number of people) hear about it?

Example: we don't own NY Times, or other sites, yet if they were to publish about scikit-learn (or in my case, Data Umbrella), it's incredibly valuable.

I recommend checking out these videos:

• [How Data Scientists Can Contribute to Wikimedia Projects] (https://youtu.be/mrHp3wc_6DQ)
• [Editing Wikipedia: Because Someone Has to... ] (https://youtu.be/s5kUSMUaZLM)

[laurburke]

I have two suggestions to add onto this: - We can look into the pages of other open source packages and tools to see the average length and content. - We should also search Wikipedia for any pages that reference scikit-learn to ensure that all of those references are blue-linked. Improving this could also raise our placement in related search results. Lauren

…

On Nov 8, 2022, at 8:50 AM, Reshama Shaikh ***@***.***> wrote:  Wikipedia is legit and an open source project. The number of users is staggering. It is considered a credible source by many users, including Funders! Also, it's super important to keep in mind that people who make funding decisions or doing research on potential projects to fund are not often users of the tool, and so even if scikit-learn is the number 1 python machine learning library, not everyone knows that. (Another reason why open source projects suffer from lack of funding.) scikit-learn doesn't need to own the page to utilize and benefit from it. It's referred to as "earned media", which is quite beneficial. scikit-learn is 10+ years old and our blog is < 1 year old. An important question to ask if how would everyone (or maximum number of people) hear about it? Example: we don't own NY Times, or other sites, yet if they were to publish about scikit-learn (or in my case, Data Umbrella), it's incredibly valuable. I recommend checking out these videos: [How Data Scientists Can Contribute to Wikimedia Projects] (https://youtu.be/mrHp3wc_6DQ) [Editing Wikipedia: Because Someone Has to... ] (https://youtu.be/s5kUSMUaZLM) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.

[GaelVaroquaux]

- We can look into the pages of other open source packages and tools to see the average length and content.

That's an excellent suggestion. I would start with tensorflow https://en.wikipedia.org/wiki/TensorFlow and pytorch https://en.wikipedia.org/wiki/PyTorch as I keep seeing scikit-learn being compared and contrasted with these.

[reshamas]

In the meantime, we should also add any Press links that people know of in a list here to add as references.

[laurburke]

Alright - I've gone through any and all pages that refer to or link to scikit-learn. I corrected any missing links on any referencing scikit-learn, updated any sklearn references to scikit-learn, and corrected the spelling to all be aligned with the lowercase name.

There were a few missing links on some pretty high-visibility pages, hopefully that addition will be helpful.