Have you ever wondered how your country is doing in terms of new players? Have you ever tried yourself to check in the European Go Database (EGD) how many dan players played in your country in the 2000's (and ragequit that quest because that was time consuming and tougher than expected)? Or perhaps, you've contemplated more than once the statistics page on the EGD website and wished it was "a little bit more"?

Allow me to bring you a sigh of happiness: A new tool appeared in the European Go Community, "EGD insights", developed by Matt Marsh (UK). 

If you had not heard of it yet and can't keep a relaxed stance until you see, you can find the website here even though it's a little upsetting for the interviewee and interviewer.
Otherwise, you can read what follows and clic there for later (and yes it's the same link).

In short, this website gathers data from the European Go Database (EGD), allowing visitors to play a little with it. That's basically the one line description you could use for it*, but...that website is much more than that!
*truth be told, it could also be described as "the EGD stats website you've always dreamed of", but that's kind of a spoiler, or more officially as per Matt himself "EGD Insights provides visualisations of data from Go tournaments run in Europe—specifically based on data from the European Go Database (EGD)", but that's definitely less catchy, right.

To dive deeper into "EGD insights", we atari-ed Matt with questions.

Hello Matt, who are you? What's your Go story?

Hi, I’m Matt Marsh, a 3 Kyu Go player from the UK.

I was introduced to Go one day when there was a power outage at a company where I was working. Since we were unable to work, two of my colleagues took out a Go board and started to play. I was intrigued, though I didn’t take up the game myself until a few years later—but I guess it had planted a seed in my mind!

In addition to playing the game, I’ve also been involved with various aspects of organisation within the British Go Association for a number of years—right now I handle rating submissions for the UK amongst other tasks.

What motivated you to create EGD Insights and how did you achieve it?

I think it’s interesting to look at trends in Go playing, and there is a lot of data in the EGD that can help with this—but it hasn’t always been possible to see those trends easily. By looking at trends we can use the information to help inform decisions being made by organisations across Europe and hopefully grow Go further.

Earlier this year I carried out some analysis on data from the EGD related to UK players and tournaments, using just a spreadsheet approach. People seemed to find it interesting, and so I thought that it might be worth putting together a website that would provide a more permanent resource—and why not include the whole of Europe in doing so? And so EGD Insights was born. Plus, it means that I won’t have to recreate all the spreadsheets each year!

Another motivation for me was to use some new technologies that I’ve not used before. In particular, I wanted to experiment more with using the latest generation of AI tools. The software for the EGD Insights project has been 99% written by AI tools.

Do you have any plans for further implementation of features ?

I’m definitely keen to get feedback from people looking at the website and I do have some ideas for additional improvements.

I think firstly and most importantly there may be other data that I’m not currently visualising that people may find useful, so I’m open to suggestions on what other charts & visualisations people may find useful or interesting.

There are some small changes which could also help with analysing the data. For example, it might be helpful to allow filtering out of the major European tournaments (like the EGC etc.), since these have a massive impact in each country when they happen—which in turn can obscure overall trends.

One larger feature that I’m thinking of implementing is to add “AI insights”. This would allow the use of AI to help you understand trends and point out (hopefully) interesting facts regarding data that you’re viewing.

Rumors have it that it opened some debates in the UK, could you say a few words about that? What lessons do you think can be learnt from looking at the data?

Seeing the data and trends has definitely opened some debates. Like all countries, we were hit hard by the pandemic and it has taken some time to recover participation. We can see that the number of tournaments has been improving—though one thing that became clear is that we still have significantly fewer Class A tournaments than before the pandemic. Perhaps this is something that we can improve on.

Another interesting fact that we’ve noticed is that around half of players in the UK only attend one tournament per year, and two-thirds only attend one or two tournaments per year. Perhaps if we had more tournaments (and therefore hopefully more within easy reach) people would participate more? Of course, this shows that the data isn’t enough to give all the answers—but it can spark debate, help with formulating hypotheses and hopefully lead to positive changes in the future.

As you probably have explored data from many countries, are there some interesting points to look at? Do we all follow the same trends? Anything striking in particular?

There are for sure some trends which we share in common (impact of the pandemic for example of course), but also areas where there are differences. Romania seems to do a great job at getting new players participating in tournaments at the “novice” (≤20 Kyu) grades. Compare that to the UK, for example, where in recent years a lot of our first-time tournament players have been Dan graded.

There are more interesting facts and trends everywhere you look, and it would be great if one of the things that comes out of this project was to look at where countries are being successful and try to replicate that elsewhere.

Would it help to add for example per country the moment when Hikaru no Go was translated/became popular, or some key historical points, for example the arrival of a very strong player from Asia, the opening of a Go school etc...and if yes, how could we contribute?

I think that adding key moments in a given country could absolutely help people to put the data into context and consider the impact of those events. Did Alpha Go have an effect on participation for example? Did the release of Hikaru no Go in a given language help? Feeding all of this information into AI along with the data might all help lead to more powerful insights.

I’m definitely keen to listen to feedback like this when people give EGD Insights a try—I encourage people to get in contact with their ideas, whether it be new features, or desire for particular charts to be added or whatever else. I can’t promise that I’ll be able to implement everything, but I’m very happy to discuss what might be feasible.

Do you have any other projects that you’re working on?

Yes!

The EGD Insights platform has two parts to it: the website itself, producing all the charts and visualisations etc., and then a back-end API to provide all the data. One thing I would like to do is to make the API public, so that people can use it to perform their own analysis of data beyond what I’ve produced myself. I’d be keen to hear from people if they would find this useful.

Another project I’m working on is a small mobile app that I hope to release soon. This will allow you to view your own EGD rating and those of your friends—and to be notified when they change. Keep an eye out for this soon!

--
Thank you very much Matt, we're looking forward to seeing all the developments of your works!

Take a look at EGD insights!