Saturday Night Data Party: Discord

All right, welcome to another Saturday Night Data Party. This week we are looking at Discord. So if you recall last week, what we did was we looked at Slack, the Slack uh communities as judged by um SEO software. So uh to recap, we went through last week and grabbed from the RF's SEO software, but again, you can use any uh SEO software you like that can track inbound links to a domain, and we plotted out just to try to determine you know how many of these things are there, how many of these things were there. By the way, if you are watching, uh please drop a comment in the comments.

That's the only way I know that you're here. Otherwise, I assume that it's just me talking to myself, which is perfectly normal. Um so we brought in the all of the known inbound links to the slack.com domain because we can tell from those uh domains what are the most uh the largest communities. And what we looked at was uh how large are they, uh what are the terms people use to describe them, and uh are there ones that were specific to marketing? So let's take a quick uh look here.

We have these uh 19,000 communities, and we sort of by traffic, uh feedly, soylent, and so on and so forth. A lot of technology based communities, no surprise there. Uh the one thing we didn't do last week that probably should is uh to to look at a timeline. So let's go ahead and create a simple plot here. Let's do uh let's see.

So we'll clean slack. And we want the first scene. And we want uh do we want anything more than that? You know, I think that's good. We'll do that, and we'll add in a count.

And we'll group by the date. And then we'll sort by the date. If we run that, we should get ourselves a nice little table here. Looks good. And let's turn that into something that's a little more useful.

Let's turn that into a line plot. We'll take our data frame here. And we'll put our count on the y-axis, put our left axis, and there's our line chart. Looks nice. Let's go ahead and copy that to our clipboard and paste it in so that we can render it.

Now you'll note that it is a little bit hard to read. There's a lot of uh you know jaggedness there. Uh, some ways you could make that a little bit easier to uh view would be to turn it, for example, to a simple moving average. Uh so let's do that. And we'll turn this into SMA of count and equals seven.

That turns it to a simple moving average, and now instead of count, we're gonna plot the simple moving average. And that gives us a little bit more uh of a graphical chart. Now, a lot of the times when we're doing code, people will say, well, you can just use this in Excel, right? You can just do this in uh tableau, and it's absolutely true, and there's absolutely nothing wrong with those tools. They are great tools.

What is important though is that when you write code as challenging as it can be to get started, uh you can then reuse it very, very quickly and not have to repeat all the steps that you just did because again you've got the code right now. So I'm gonna take this and literally going to copy and paste it. We're gonna start ourselves a new document here, and we call this, and we will call this Discord instead of Slack. Right? So we're gonna change our data folder from data to Discord.

Right, and then just do a very quick fireplace just for simplicity's sake, so we're not overriding uh data frames. Anyway, wherever we use the word slack, we're gonna use the word Discord. And now let's see what we've got. Because again, these are the exact same data files that now from the same tool, so we should have the same format. Let's see what we got here.

Yep. So already we've got uh Slack brought in. Hmm. That's exactly right, Chip. Uh we are we are exactly reusing the previous stuff and doing it uh with Discord's data instead.

I'm just gonna double check as one should always should to make sure that your the data is in good condition here. That's what I thought of. We've got pulling from the wrong data folder. I knew something did not add up. Okay, so our Discord data frame is 296 279,000 observations.

There's our traffic. Now let's go through and something through an error here. So we need to figure out what threw the error. It did not like traffic. Ah, it's because we didn't do a type convert.

It is uh still throwing traffic as a character. Alright, we're gonna have to course that manually. There we go. Alright, now we can go through step through this and see how it looks. Still throwing some still think we're still throwing it off because you can see there's only 33 rows there out of 290 279,000.

There's the force. What's ah Discord does not have separate domains. When we look into the into the domain field here, the link URL, it's all Discord.com. See on Slack's links are all, you know, like uh trust insights.slack.com. Discords are all very different, uh they're structured very differently.

So you don't see the um a change in the name of the community. So if we pull the domain name on this, of course we're gonna get very, very few. So we need to actually remove that from our code and see what we get here so we've winnowed down to our eleven variables into now we have a a bit of a problem still right we have all of these discords but we don't have distinct domains so we'll probably have to do just a unique on the link url itself so let's do that distinct link url keep all as true now how many distinct discord links do we have 167 thousand that's a decent number and we're sorting by traffic so let's take a very quick look at our top discords porn hub no surprise there if you notice too we've got some we've got instances where we have two different um discords right with different links the older ones and the newer ones but same similar page names so let's do let's do a unique on the referring page title and just copy this entire line page title. And now if we do a unique on that, we should be filtering out the duplicate pages. 114,000 now.

Uh yes, Chip. There's uh the first year as a preview day to show how different it is from Slack's topics. Yes, yes, indeed. Even just uh we got online gaming, we've got uh more porn, uh Fortnite, uh new eggs at Discord, uh online chess OBS, which is the software we're using to stream this, um, more porn. I'm getting a I'm sensing a theme here.

Alright. Let's write that out and take a look at. Chip says you didn't realize there was a new egg discord server. Yes. There's uh a ton of these uh things here while we're waiting for that to render.

Uh see League of Pool Beach Party, Nexus Mods, which I'm uh Android emulator, paper.io, WoW Head, which is the World of Warcraft stuff, super smash flash gaming, NBA streams. Watch NBA. So there's some interesting communities in here. Um some of these, you know, like obviously some of these are not safe for work. Um Wow Progress, which is uh they have Minecraft, more porn, more porn.

Uh I'm surprised that there aren't more um cryptocurrencies. So there's still obviously a very heavy gaming focus on stuff, but I'm honestly surprised to see um not see uh more bit more like you know bitcoins and cryptos and other stuff in here, uh cracked apps still more porn, dungeons and dragons. Let's get a look at some of the the numbers too, out of curiosity here. So traffic uh let's see twitch's Discord, two million visits mobifier, half a million. I mean these are substantially larger numbers than some of the slack ones.

The slack ones, you were talking a couple of thousand visitors. Um this is our so this data should be monthly. Uh, here you're talking hundreds of thousands of visitors. So off of the top of my head, if you are trying to reach a large audience around a specific niche, um, and it's consumer focused, particularly gaming and entertainment focused. This is definitely the platform uh to be spending a whole lot of time on.

How far in are we here on this tokenization? Ah, all right. The tokenization is done. Let's see what we got. We have join Discord, invite links, voice text.

Okay, so nothing's super nothing super stands out here in terms of the names. Let's take a look now at um our see if any of these are there any marketers in the house. 39 Discords. Let's see, blog traffic. Uh burst report.

Oh dear. Um traffic numbers on these almost zero. Okay, so clearly if you're looking to host a marketing community, uh, there really isn't one uh here to speak of, so we I think we can safely rule that out. Uh let's take a look at close some of these frames. Timing and run this.

That's a lot more sudden. So uh let me pull back up and rerun Slacks so that we have we can do side by side. I don't need the tokenization, I just need the the raw chart. Okay, so there's slacks, and you can see Slack um seven-day moving average, number of visits. Uh you start picking up 25, 50, 75, 100.

Um over time, and then we go back to this plot with Discord. Um, go from essentially crickets until like really recently, pandemic would make sense. Um thousands of visits a day. Oh we um a month. But these are I'm sorry, no, these are not visits, these are in number number of servers.

Yeah, these are unique instances of server numbers because that we're using the first scene date when when RFs detected the the inbound link so you have a huge influx of new communities again pandemic no not a surprise there right this would be the first right here is the halfway mark of 2020 um and right here is about the the quarterway mark people created a lot of discord communities whereas in the slack world um you're talking about a hundred new communities uh in a uh a uh uh day in the same time period so discord's huge but obviously very disparate um if we look numbers wise we clean these out we had 1900 slack communities that were detectable that had were unique in this time period had 11400 discord communities it's bigger by uh in order of magnitude uh and it's happened much more quickly so if you are marketing and you're thinking about where you want to build your house if you're doing it B2B I think Slack is probably the place to do it right there's not a ton of competition um there's obviously not as many new communities being created each day although the uh these curves definitely are are eyebrow raising if you're B2C in gaming entertainment of any kind it's gotta be discord and you're and there is more competition for eyeballs it is huge um and there are um there are some neighborhoods that understand that you're gonna get there's a lot of people on them, but there are some topics and things that your company may not feel comfortable um I'm about to say being next to because all Discord servers are are functionally independent. They don't like cross over each other, but just know that the audience overall has a many more varied interests on Discord than they do on Slack, just from looking at uh the names and the topics of them. Um if you're interested in seeing more of this data um or you want to talk about it, please join our free Slack group. I know we've we now look so lame because you know Discord is so much bigger. Um but join our our Slack group.

Go to trustinsights.ai slash analytics for marketers and talk about this, look at some of this data uh questions you have. Thanks for folks who are uh on tonight, uh like Chip that had some uh some fun questions and things, and uh we'll talk to you soon. Take care and thanks for watching. Want help solving your company's data analytics and digital marketing problems? Visit TrustInsights.ai today and let us know how we can help you.