RB Search Gizmos

Find Relationships Between Wikipedia Entities with Crony Corral

Do you sometimes feel that you should have a pen and paper for taking notes when you read the news? Do you long for a program that lists all the people involved in a news story so you can get details when they pop up after an extended absence?

It’s hard to keep up with all the current events when there’s such a ferment of things going on. It seems like your choices are either to severely restrict your news intake for your mental health, or go around in a state of half-understanding which certainly isn’t good for MY anxiety, I don’t know about you.

Even when we’re not consuming the news, though, it’s still being made. Newspapers are still publishing. And Wikipedia is still churning away, aggregating and integrating news content, as people far bolder than me try to make sense of everything.

I wanted a way to harness Wikipedia’s consistency in keeping up with news and turn it into a tool that would allow me to examine the relationships between people outside of a single news article. And, after several false starts and multiple epic discussions with Curly, I am pleased to present to you Crony Corral, at https://searchgizmos.com/crony/ .

Crony Corral accepts input of names – people names, organization names, or company names – separated by commas. Once the names are entered, CC searches for them on Wikipedia and pulls their Wikidata properties, looking for matches across 17 different Wikidata properties:

  1. P159: headquarters location – Location of the main office of an organization, company, or institution.
  2. P108: employer – Used to link a person to the organization or company they work or have worked for.
  3. P69: educated at – The educational institution(s) a person has attended.
  4. P551: residence – The place where a person lives or has lived.
  5. P102: member of political party – The political party a person is or has been a member of.
  6. P106: occupation – Refers to the main job or profession of a person.
  7. P39: position held – Used to link a person to the political, organizational, or professional positions they have held.
  8. P937: work location – Indicates the place where a person primarily conducts their work.
  9. P452: industry – Refers to the main industrial sector or sectors that a company, organization, or product is involved in.
  10. P17: country – Indicates the country that a geographical entity or organization is part of or associated with.
  11. P1056: product or material produced – Refers to the main product(s) or material(s) produced by a company or organization.
  12. P749: parent organization – Indicates the higher-level organization that a subsidiary or lower-level organization is part of.
  13. P414: stock exchange – Refers to the stock exchange where a company’s shares are traded.
  14. P112: founded by – Indicates the person or organization that founded a company, organization, or institution.
  15. P127: owned by – Refers to the person, organization, or entity that owns a particular asset or resource. (I think institutional / stock ownership is in here too.)
  16. P355: subsidiary – Used to link a parent company or organization to its subsidiaries or lower-level organizations.
  17. P27: country of citizenship – Indicates the country where a person holds citizenship.

When matches are found, they’re gathered into pairs which are presented to the user in a drop-down menu. Users can choose a pair from the menu and find Wikipedia pages which the two names have in common with a relationship level of 1 (minimal relationship) to 5 (close relationship).

Here’s how it works.

Using Crony Corral

Screenshot from 2023-04-10 09-44-03

Start by entering a list of names. They don’t have to be all people names or all company names, but only people will match people and only organizations will match organizations (companies, NGOs, etc.) That’s because the two groups use different Wikidata properties. I have put in 75 names at a time without a problem, but it took a good little while to sort them all out and the menu of matches was huge.

Let’s not put in 75 names. Let’s instead use the very basic example of Louisa May Alcott and her father, Amos Bronson Alcott. Put those two names in (separated by a comma) and click the button.

Screenshot from 2023-04-10 11-57-09

After a moment the menu of sub-groups will populate. This shows all the matches from the group of names you entered. Even though we only entered two people, there’s still a long list in the menu. That’s because being family members, Amos Bronson and Louisa May Alcott of course shared many life experiences. You’ll note that they match for several residences. That’s because Crony Corral considers all values in a Wikidata property – all employers, all occupations, etc – when looking for matches. Which of these menu items  you pick, however, doesn’t matter, because they all represent the same two people.

Screenshot from 2023-04-10 12-01-42

Once you’ve chosen the pair you want to review, you need to choose the minimum mentions threshold.

What happens in this next part is that Crony Corral searches for Wikipedia pages that the two names you’re looking at have in common. It then counts the number of times each name is mentioned in that common group. Pages with only one mention of each name would be the slightest-possible affiliation, while pages with both names mentioned five times would indicate a close association.

Screenshot from 2023-04-10 12-11-25

If you want to see all mentions, set it at 1, but if you’re searching for famous people pairs you’ll get a LOT of results this way. Here’s what Amos Bronson and Louisa’s common articles list looks like with a minimum mentions setting of two:

Screenshot from 2023-04-10 12-21-31

There were nine results, covering a variety of people, places, and things. If you’re an Alcott fan some of these pages might make immediate sense (Fruitlands) or they might take a moment to put into context (I always forget the Alcotts lived in New Hampshire at one point.)

Each article has three search links under it: one for Google, one for DuckDuckGo, and one for Bing. The search links search for the title of the Wikipedia article as well as the name pair you’re searching for. If you clicked on the the Germantown Academy (not shown) Google search link, this would open in a new tab:

Screenshot from 2023-04-10 12-30-44

These search results get right to the point: Louisa May Alcott was born in Germantown and that’s why the two Alcotts have a relationship to that Wikipedia page. And do you see how rich the results are, how focused they are on information? That’s because you’ve added the additional context of Germantown Academy, which is getting you past shallow SEO and ecommerce results. There might be all kinds of companies trying to rank for Louisa May Alcott or even Bronson Alcott as a search query. I don’t think anybody’s worried about ranking for Louisa May Alcott Bronson Alcott Germantown Academy.

What happens when you want to explore a network around someone but you only have one name? Use a different Gizmo first: Wiki-Guided Google Search ( https://searchgizmos.com/wggs/ .)

Getting a List of Names With Wiki-Guided Google Search

Screenshot from 2023-04-10 13-04-36

Sometimes you’ll have one name to start with and you want to explore the network of that one person – perhaps you want to explore Amos Bronson Alcott outside the context of his family. How do you get started? Use Wiki-Guided Google Search. It’s designed to find you related pages for a Wikipedia topic, so this is a little off-label use, but it still works. Set your topic search for Bronson Alcott and your minimum mentions to 1. You’ll get a long list of results, including a lot of names.

Screenshot from 2023-04-10 13-04-07

Going through this list got me over a half-dozen names, which I added to a Crony Corral query along with the original Amos Bronson Alcott search term. From there I have plenty of intersections to explore.

Screenshot from 2023-04-10 12-50-31

2 replies »

  1. I would *love* to read about how you are conversing with Curly about the programming work. That seems to be an entry point to building a skill that looks valuable.

Leave a Reply