Hello, and welcome to this WikiCite session. We're going to talk about Swedish parliament and judicial documents and their use. And your host tonight is me, Jan Einarle, and Daniel Einarle. Hello. How are you doing, Daniel? Very well, thanks.

So it's a wonderful, rainy evening here in Stockholm. All right. We had a pleasant day here, a little bit first of an autumn day here in Amsterdam. So what we're going to talk about today is quite of a long project. But we're still just in the midst of it, I would say.

We have so much ideas. But it's about this golden treasure of the open data portal of the Swedish Riksdag, the parliament of Sweden. But before we go into the depths of that, we want to give you a little bit of introduction. We have got started.

And before we go into depth, I want to give you some sort of housekeeping things. So everything today, we have the WMF, the Wikimedia Foundation Friendly Space Policy that we're here to. So please be friendly to watch each other in the chats. And we have the chats activated here,

both on YouTube and on Periscope. And we're also streaming this to the Swedish Wikipedia group, Facebook group. So you can all comment there. And we will see it and can bring it up. And we have already had a friendly hi, all from Fili here. So please ask us questions during this entire session,

because I think there are things that you might want to know more of that we have so internalized. We don't even bring it up, because we have been thinking and talking about this a lot. Yeah, there might be a lot of special peculiarities with how the Swedish system works, but we don't quite

think about it, but not everyone knows. Yeah, and we have learned a lot about the Swedish peculiar system when we have digging into this. And I think that's almost like it's everything when you go into Wikidata and think, ah, I know this topic. I can model this. And then, of course, you see, ah, no, it's

much more complicated in reality. And what else do we have? So you can comment directly where you're watching. But if you rather want to, you can also comment in the Telegram group for Wicca site.

And yeah, that's about it. So let's dive into the introduction of Swedish project. I'm going to share my screen here. This is the share screen.

So WikiProject Sweden, I guess most of you people know about WikiProjects in general. And Wikiscientist, of course, kind of a WikiProject. But when we started here, it turned out

that there were already a lot of things done in regards to Wikidata and Sweden. And I guess most of the countries have these kind of properties, country box, where you can see the properties that

are available to you and related to your country. And that's sort of where we started. But when we started to get into the nitty gritty of things, we saw that there were already things being done before. We even got here. So one of the first thing is this subproject

from the WikiProject, Every Politician, which is, of course, a much larger project that started years ago and which are much larger scope, like trying to get every politician onto Wikidata.

And see if it loads here. Yes. And we have this little bit of tracker here. And I think we have just gotten up to two stars here on Sweden during this, because one of the first things we wanted to do before actually going anywhere was to get all the parliamentarians into this. And we're going to meet some of the people who

have been in that broad process as well. And the most complicated things about that was that most of the politicians already had an article on Swedish Wikipedia, but their Wikidata items weren't very good at all.

So we had to improve them. And that was tough, because since they had some data, it was hard to run a bot. Because it's easier if there's no data. You can just add everything. So we did a lot of work by hand to get that up and running.

How did you solve it? Oh, I think we may have just lost Jan. And I think I just exited the stream there. Good to have you back.

Yeah. I was trying to ask what you would do to solve it. Maybe that's going a bit or too fast into what we were going to talk about later. How did you solve the problem of disambiguating or not adding duplicate politicians? No, I think that's a great question to start with,

because that's what you run into in the beginning. And luckily, in the Swedish open data set, there's an identifier for the parliamentarians, which has been of great help to actually point like this is exactly this one.

Because, of course, you have a few parliamentarians with the common Swedish names. So you can have three or four with the same name, but they've been there over the years. So that helped us a lot to get started. And then let's see if I can get back to share my screen here again.

This is where we were. Yes. So that's sort of where we got started. And this actually started last year about this time. And I thought, I can do this by hand and be done by Christmas. And it took a little bit of a longer time than that.

But eventually, we got it done. And then it was easier to backfill with the rest. And then when we had the politicians there, it also seemed easier to start adding more things to it. And that's where we actually got into starting a Wiki project.

And of course, we weren't the first trying to do a Wiki project for a country. So we copied proudly the design from Wiki project India and just customized it with our colors and made a logo for ourselves.

And that's in our background here. And then we started with some plans and ambitions, like what are we going to do with this project and how can we collaborate? And a few people joined. And we actually started a Telegram group just so that we could chat in Swedish and not spam the general Wikidata group

with all our nitty-gritty details, because there are always a lot of nitty-gritty details. And then when we had something like, oh, this is a larger question that relates to more countries or all the countries, then we bring it up to the general Wikidata group or going on Wiki

and having a discussion or going to a property talk page. I think having a place to shout out ideas or questions in a sort of more alive fashion almost, or having fast, real-time answers is very helpful when you're stuck on something and you need advice.

So it's been a very good catalyst, I think, for getting some of this work done that we're going to talk about today. Yeah. And I'm going to mark a few of these. These projects were already related to Sweden and existed before we started this project. So there had been subprojects in specific topics, some of them

quite large, because the Wikilabs monuments, that includes, like, 150,000 monuments. So that's quite a big one. You know, they're working with these things, right? Yeah. But this one, now we could actually join forces,

everyone who had some sort of common thing about Sweden to some interest. And these are the two things that we're going to talk about today. We're going to talk about the Riksdag documents

and the court decisions. And of course, here are all the things we have. Do you have anything, thoughts in general, Daniel, on how this overviewing project has helped us?

So I think, I mean, it's been a place to sort of showcase, I guess, what is happening around Sweden-related topics in general and to bring together sort of, you know, I would say dispersed group of projects that have some kind of relationship to Sweden

that we're a little bit overall on Wikidata. And I think it is mainly, I think, this is working as a showcase where people can get ideas for, oh, someone has done this already, but what if I were to do something similar but with this other data set or without this other group of objects? Yeah.

And then, of course, we have had this project page to talk on. We've had this Telegram group in Swedish, which makes it very easy for us because, as you can hear, we're not native English speakers. This is our second language. So as soon as we go into details, especially when you go into details

about the parliamentarian processes, like, it's tricky even for a Swede knowing the difference of the inquiry reports and other kinds of documents and what they actually are. So falling back into your native language has helped me a lot.

Yeah, I mean, we're trying to keep the discussions on Wiki in English to, I guess, to share our thoughts and ideas with other people who might be interested in seeing them, but having a place off Wiki where we can discuss that in our native language has definitely helped. Yeah. And of course, doing it on English on Wiki

has also helped because we have gotten a lot of help from other people discussing how we should use different kinds of properties. I think one of the biggest thing was that after getting all our parliamentarians on there, the way we connected them and their offices

to the party they represented, we used the wrong property for that in the beginning. So we used the member of party where we should use the parliamentary group and the nuances in the labels in Swedish made it really unintuitive for a Swede to know this is not applicable to us.

And then after discussing like, no, but every other country using it in that way. So perhaps we should just change the label in Swedish to make you understand. And of course that makes querying a lot easier afterwards. And I think we will get back to that a little bit later how that is, of course, a factor in everything.

So just quickly going through this project before we move into the other things. Yes, so what I was gonna say, besides the telegram group being on Wiki,

we have also doing online edit-a-thons, not really edit-a-thons. We have more like having online meetings to discuss because we haven't edited much, have we Daniel, during this? It's been very little editing. Yeah, but we meet like having a steady time each week,

just sitting down and chatting a little bit, well, like what happened and what are ideas. Sometimes that unlocks so much of the things that are hard to express in words, especially when you come into the complex things of qualifiers and modeling.

And I think also, even if the chat is great, I think talking, well, not face-to-face, but at least on a video meeting that you can maybe faster iterate or come up with with new ideas faster in that setting. Yeah.

The monster. Yeah, so let's head into our first, like topic of today, which is, is it a court decisions we're starting with? I believe so. Yes.

Do you wanna share the screen? Yeah, I think that might be a good idea, but perhaps I can give some introduction just to what it is, you know, without sharing the screen. Mm-hmm. So, hey, so why would you wanna do this? Why would you want to put court decisions on Wikidata?

So Sweden is a bit different from some countries that, well, the one that I know mostly, perhaps is the US, which you hear a lot about in the news because the Supreme Court in the US is rather powerful and important. But the Swedish Supreme Court is maybe a bit more anonymous, but it's still quite important

and has in recent year, I think, years, I think, taken a bit more of a, or using its power and actually setting precedent where before they maybe would not wanna step in where the law is unclear or ambiguous. Now they actually step in and say, this is what we believe the law should be. And if the legislative branch is not happy with that,

they will have to change the law. So they are a little bit changed the way they do things. And it's kind of interesting to be able to look at some of those patterns, right? And of course, what I looked at was as well, the US. So on Wikipedia, pretty much every Supreme Court case,

I think in the US has a Wikipedia page already. Everything is modeled great in Wikidata perhaps, but there are some of them that are modeled out. So I'll show you one of my, or one of the more famous perhaps US Supreme Court cases.

Have you already put my? There it was okay, yes. Yeah, so this is quite a famous case, right? The Citizens United versus Federal Election Commission, which is, I think about, probably some American is gonna tell me I'm wrong here, but I think it's about how companies can finance campaigns in elections.

And specifically it said that corporations do have the right to free speech and financing a campaign in elections is free speech, according to US law, or the US Supreme Court anyway. So that's kind of what this case said in a nutshell.

And if we go through quickly what it is, maybe I should change this to- To English, so yeah, how are viewers perhaps? Yes. Indeed, of course it's been an instance of a Supreme Court decision. For some reason, there are two instances out here.

Maybe someone should look into that. Was the difference between a Supreme Court case and a Supreme Court decision, I'm not sure. So it has a country and the jurisdiction. It has a publication date, obviously. It has some significant events that are all the oral arguments and the decision dates.

It's got a citation and this is how you would, if you write a text about this decision, how would you refer to it? So this is getting into the interesting stuff for WikiCite. And there are often for legal texts and court decisions, there are ways to refer that are rather standardized. And in the case of the US Supreme Court, there are a number of ways

you can refer to a court case depending on sort of which I think publisher you are using for your, I think there are, these are duplicates from it. Probably, yeah. I should probably. No, they're different. There's a space in between there.

Is there a space? Ah, okay, maybe I should not. You should leave that to someone else who knows better than I do. So it has a court, right? And then it's published in the US United States reports. It's got even a common category. It's quite a famous decision.

It's got a majority opinion that was written by Anthony Kennedy here. All right. Who was that? That also has the reference there. Indeed. So I think the Supreme Court did this. And then there are a few other identifiers in the page on Wiki.

There's also quite a lot, if someone is curious, there's also quite a lot of other information on the Wikipedia page that you could actually bring in and insight on there. For example, there are a lot of opinions on this that we're gonna talk about a bit later. How we actually modeled the opinions.

Yeah, I think the PBR once that had us here. Indeed. No, so that was a US example.

And so I looked at that and stole quite a lot of the modeling. I guess there's one more thing that is up to actually showcase on this item, which is how you model the different opinions. Then I went to sort of, okay, what can we do for Sweden here, right? So what is the data source for this?

Well, of course there is the Supreme Court website and this is all in Swedish, but let's open up one and look at it. Anyway, I think I'm gonna search for one that is from last year's because it makes life a bit easier to show something.

So these are all the different decisions, right? We can maybe pick one, the funny name perhaps. Trying to see if I found something interesting,

but maybe not. It'd be about a leaking roof. The leaking roof, all right. Sounds very interesting. So this was a question about the responsibility for correcting a problem in an apartment where there was damage posed to the apartment

and to another part of the house. So who was responsible for paying what? So typical litigation around who pays what and probably some insurance companies involved here. So this is the sort of website that I had to work with when it comes to sources. And this is the website and then you have a link to the PDF

which contains the full written decision. And of course for 2019 and I think a few years back, you have a little bit of actual data here on the side, which helps. So there's the law supplied, there is a name that is given by the court to the case, there is a case number and there's a few key words that are sort of helpful.

So one of the first things I did was just to sort of scrape this website and scrape the name. And this is what we talked about, the legal citation. So this is how you would refer to this case in Swedish legal text, as the legal scholar. But then pretty soon,

I realized that this information is not that great. So we actually have to go and look at the final case. Now you're not sharing that view. How do I share the PDF? You'll have to share again.

Stop sharing and share a new one.

So here it goes. So here's the written case then. And it's got some interesting information here. So one of the things that I really wanted to get at was, it's actually down here, very bottom.

It says in this decision have been taking part that justice is So these are the judges that took part in this decision.

So this is a bit of a Swedish court, perhaps. Not like in the US where the whole court always decides in the full chamber of nine justices. In Sweden, there are typically five justices that decide one case. And the court has, I think, 16 or so justices,

and they are specialized in two different departments, one about criminal cases and one about civil cases. They tend to only decide decisions in the cases within their area, except for a few very special decisions where they overturn a part of previous precedent

where they would actually have a full chamber of all the 16 justices that decided together. And then there's another peculiar thing here, which is the called referent in Swedish or maybe a reporter or rapporteur or reporting judge in English, which is the justice

that is in charge of that case. And so that's also something we wanted to record. And then try if I can be a little bit faster this time.

So if we go back then to actually look at this case, we can search for it. So the name is available on Wikidata. The search for it is the name of the item, of course. And so here we can see what the decision

from the Swedish Supreme Court would look like once it was imported then. So we have it, it's a decision of the Supreme Court of Sweden. It's got a title, which is set by the courts. So we consider this the official title. And if for some cases that are older,

that was not given a title by the court, we just put this as no value to indicate that. So that way we don't get, for example, problems when we add citations to these documents. Otherwise we would get a constraint violation because the document doesn't have a title. It's got a language, a publication date, the citations.

So some of the things you've seen. And then I've added here, which are the judges that were part of deciding that case. And we've added a little qualifier here for the reporting judge. I guess published in is kind of interesting. It's published in a special volume

of a yearbook of this court that is called Meturidis Dalki. So we can actually, these actually have ISBNs and stuff. So they are modeled like a typical edition. An edition object. You have a link to this PDF as well.

We have a majority opinion here. So this has part majority opinion. And here we come to the different types of opinions. In this case, all the five judges agreed or justices agreed and they all ruled the same. So they all joined this opinion. There's not really information

about who wrote an opinion in the courts. We don't have an author on these in Sweden. And then we have added some citations here also, which we're gonna talk about much more in depth how we did later. But these are citations to few items, which I think probably represent other legal texts

that we're also gonna talk about later. So the actual draft legislation that was going through parliament, for example. We also got a special Supreme court case number, which is a special identifier for all these cases. We can use to get back to that website

we saw earlier on the court. And that was a brief overview. John, do you have any questions? Do we have any questions from the audience or anything? We have some general discussion about court cases and how they could be modeled in like the case and the decisions. And to be clear to us,

then we have mostly been working with the actual decisions in this import. And you were about to start with that, but what we were starting to dream about why start with the court decisions is that we realized that since they are so important,

they start sort of a chain reaction. Because the Supreme court is the ruler of how a law is gonna be interpreted.

And then it's might lead to politicians wanting to change the law. If it doesn't go in the way that you thought it would be or how it was supposed to be or times has changed since the law was written, perhaps by technology or society.

And here we started to imagine like, all right, so this could be the start which feeds into a political process. And this is where the parliamentarian documents will take on. Yeah, it's exactly right. We've seen, I think a few cases in the last years now

where these, as I said, the court has made decisions where the law is ambiguous and then sort of more or less told the legislature, the legislative branch that you should fix this. Yeah, and do you have any examples of things this refer to or from?

Maybe, I was thinking of looking at this opinion so we can see. Do you mean a court case referring to something? Not citations, but. Yeah, or maybe we take that in the second session with the actual citations.

I had otherwise these sort of cause and effect things. I think the example I was thinking about now is a bit too recent. So I don't think we have that modeled, but there was just a few weeks ago in case of a court decision

that led to the government starting an inquiry into the change. This is an example of something a bit more difficult when it comes to modeling the different opinions. And this is a model that I did not draft. It was drafted for the US Supreme Court cases.

And I just stole that proudly. So we've got majority opinion here. There were two justices that agreed on the majority opinion, which is the main body of the text. And then there were two justices that had a dissenting opinion that meant that they wanted to rule differently in the case. They have a special opinion supporting that.

And then you have a concurring opinion, which another justice had, which means that they agree with the decision, but they don't agree with the reasons that were given. So they have another way to get to the same result. All right. Here's another peculiar thing here, which is called an addendum,

which is often written by, or often, but sometimes written by the justices to sort of either clarify their own position or why they voted in a specific way, or they want to tell the legislator that they think it's ought to be fixed. So that's, for example, Stefan Linsco,

who is the former, who is the former chief justice. He was kind of famous for writing these opinions to the legislator saying that you should fix this. So he wrote actually quite a lot of those. All right. So hopefully they will show up

when we're getting farther into this project. And I guess maybe we can talk a little bit about the challenges of importing. You already saw that this was a PDF file that we were reading this from. So I actually didn't want to go through

all of these hundreds of cases, about 100 cases every year, I think the court hears. I didn't want to go through all of those by hand and adding the justices. So as with the politicians, one of the first things that they did was actually just make sure that the whole history of justices on the Supreme Court was correct in Wikipedia. And I had quite a lot of help

from the Swedish Wikipedia page that had a very nice list article that I could basically scrape and import to Wikidata. So wherever justice was a red linked on Wikipedia, I sort of could almost guess that it would not have a Wikidata item either.

And that helped me get them in there and it had also the period of time they were on the court and so on. So I got quite a lot of information for free that way. So once I'd done that, I could write a little bit of a script that would go through and scrape those PDFs.

Well, it was a bit tricky because those PDFs, they aren't written to be machine readable and there's slight variations in style, how you write it. And it had varied a little bit over time. So there was quite a lot of trials and error. And in the end, I got about, I think a five to 10% error rate on them.

So I would still create most of the information about the case and the item, but there might be that the justices were not included or there was not all opinions were given, justices that joined them and so on. So it would just let my script spit that out in the log for me. And then I would sit and work through that log file

after I've uploaded to like five years worth of cases or something and just manually go through that file. So instead of having to do a hundred cases a year, I could do maybe five or 10 and just parts of them, parts of them manually. So that helped quite a lot. Now you used Wikidata integrator

and the custom script for doing that essentially. And I think we also almost skipped a little bit of a step here because you touched upon it when you mentioned modeling, but I can really recommend using some sort of a template

or a table. And there are these different kinds that you can use. Now I'm zooming in too much. I'm gonna zoom out a little bit here. Maybe I skipped that because I myself did not know about this when I did it. Yeah, so this is a place where collaboration on Wiki

is really helpful because here we can sketch out how we think we should model this based both on the data that we have and also on how we're seeing the properties being used on other places on Wikidata. And it's also a great place for other people

to find and come with suggestions. And as I said, I think I actually sketched this out partly after the fact or after I started my imports, but yeah, it was quite useful. And then we have something else that is useful below there.

Yes, this is a great tool called Integrality, which helps you keep track of data. And here we're still trying to figure out what tables are most useful for us, but we added a few that we thought were gonna be interesting, like which laws apply to these court cases?

Are they citing other work? And here you can see quite a lot of these cases actually do cite other work. And then of course, main subject, which can help you find topic areas of things.

And yeah. We still can see that we still have some work left to do on the classification. And here we see this sum so far then of the number of court cases. Or decisions. And with that, I think, should we move on to the parliamentarian documents?

Yeah, that sounds like a good one.