In a previous post, I highlighted easy searches that return a range of useful information — the kind of wizardry you cannot do on Encyclopedia Brittanica or a classical library.
With a little bit of technique, you can do more specialized searches that reveal even more insights. These queries run on SPARQL. I had previously not heard of that interface/language either, but quickly got going.
Mapping Entries
Let’s say you want to find all the places that are mapped in your country. To get this, you run a query for all the places that are listed and connected to that country identifier.
You can render this onto a map also, which is particularly intriguing to look through. Here we are looking at Kvemo Kartli, and some of the remote points are megalithic fortresses. Not all entries are mapped, not all remarkable sites are entered. (If there are fascinating sites and locations missing — add them.) You can zoom around the map, and check sites outside Georgia’s borders that connect to the country also.
The query is more complicated than those in the previous post. It looks like this:
Asaf Bartov, Grand Master of Wikidata (or so he appeared to us apprentices), taught us how to build such queries in a short workshop at the Wikimedia Central and Eastern Europe Meeting in Tbilisi, mid-September.
Via Picasso (“good artists copy; great artists steal”), Bartov encouraged us to use his code. In this case, I have tweaked it from a workshop he ran in Estonia in 2017, replacing Estonia’s “wd:Q191” with “wd:Q230”. Bartov kept repeating — “Wikidata loves you” — if you click on any of the variables and type “CTRL + Space” you can search the text variable you are interested in. I have no programming experience, but found it quite intuitive.
Click on the link symbol, and you get the link to the code — allowing you to run it again, or steal tweak it onward. Here is the link.
Images & Paintings
Want to find all the paintings that are listed as featuring Georgia? You can do that, too.
The code is here.
Armenia has more paintings — but still inconsistent in its listing. Once you have the structure for the query, you just change the country. It’s a matter of 10-15 seconds.
Here is the query for Armenia.
Biographies & Occupations
Another thing Bartov showed in Estonia and which I adapted — what are the occupations of people listed as Georgian in English & Georgian Wikipedia? Overwhelmingly politicians. Interestingly, soccer players are among the top five also. Note that these are *all* the instances, so a soccer player who starts acting and then becomes a politician might have several instances.
Next to the top 100 most read articles (see my articles here on the South Caucasus and on Georgia), this again provides an interesting glimpse, and a way to compare what people engage with across countries. The code is here.
Father-Son Politicians
Here is a query that looks at politicians whose fathers also were politicians — like the previous searches, this was also tried in Estonia.
Some are missing. This brings us back to it being a great idea to improve coverage. The code is here. You could easily tweak it for musicians or writers.
How to learn this in detail? Bartov provides a more detailed introduction in a two-hour tutorial here. In an age of instant video snippets, two hours may sound daunting — but in the greater scheme, in a short time you open up a whole new way of looking at knowledge, and how it is generated and used.