Ever since I started reviewing one game a week I’ve eternally struggled with managing my time to make that once a week post. As many of my fellow adult gamers will attest to time is an ever shrinking commodity, especially when it comes to leisure. So I started looking for games I knew I could finish in a week, something with a play time of maybe 4 to 8 hours, but the more I looked for play time stats the more I came up short. There are resources out there of course (How Long to Beat being the shining star among them) but for new games, the ones which people are most interested to see reviews of, the play times aren’t known until the games have been out for some time. So I did what any average coder would do, I started building my own solution to it.
Long story short: I’ve been collecting various bits of data from Steam’s Web API for some time now and it’s reliable enough to provide some insight into games that are released through it. Whilst there’s numerous aspects I could dive into I felt a brief, concise infographic done on a weekly basis would be an interesting exercise, one that would hopefully spur on further conversations about why games were popular, what developers are doing right and, of course, what they’re doing wrong. With that in mind I’ve spent the last month wrangling my data into a usable format and putting it into something digestible, the first of which you can see here (and the second of which precedes this post).
As with all infographics there’s a lot to talk about the data I’m presenting and this post will attempt to provide some insight into what you’re seeing, some of the decisions I made in presenting the data and why some things might not exactly line up with your expectations.
The data I’m using is all publicly available from Steam profiles. If you’ve set your profile to private I can’t see anything: not your achievements, your friends or even any kind of play time stat. That being said if you’d like to be excluded from my data collection just shoot me a message and I can ensure you’re not included in any future collection activities.
I’ve chosen to do this retrospectively as it takes around 2 weeks for most games to get good, reliable data.Thus these will always be 3 weeks in the past, giving all games that are released in that window a minimum of 2 weeks of data collection time before I make any results public. Typically this means I have a sample size of around 200,000 players to work with which I think is large enough to be representative of the larger Steam community. Indeed the few sample runs I did before publishing any of them seemed to line up with my expectations although I’ll be the first to admit that my statistical analysis skills have diminished quite a bit since my university days.
The first section contains some quick stats about the week with a comparison to the previous week’s stats. The one part which might get reworked in future versions of this is the “Top Genre” as this is just based on the number of games released in that genre that week. Indie seems to dominate this pretty much every week so if there’s another high level stat that you’d like to see included I’d very much like to hear your ideas.
The second section is the top 5 games, by total hours played (that I’ve observed), that were released in that week. This does mean that Early Access games, especially those that have spent a long time in the program, tend to stand out however I’ve made the decision to include them for a couple of reasons. For starters that does mean the game is popular and most of the time even the biggest Early Access games still lose out to big AAA releases. Additionally looking at other stats for some popular Early Access games that do eventually “release” they’re usually quite popular in their release week as well. If I find that’s not the case however I do reserve the right to remove them in favour of more deserving releases (although that hasn’t happened yet).
The map (which you can only access by visiting the blog) uses the same metrics as the previous section but at a regional level. If you’ve put your country in your Steam profile then its possible for me to see that and I can tag the play time with the region. The data I have is a little more fine grained than I’m presenting here as Piktocharts has around 170 countries listed whilst steam has around 250. There’s also a couple countries for which I don’t have equivalent data so they’re unfortunately blank. However this does serve as a good way to see what games are popular where and how countries differ from each other.
The trending section is the only one which branches out from games released that week (in fact it excludes them from this list). Trend scores are calculated using a Z Score that compares the average players in the game during that week to the month that preceded it. The higher the Z Score the higher they appear in the list. To qualify for the list the game must have attracted an average player base of more than 100 (to filter out games that went from say, 0 players to 10, which gives a massive score) and manage to maintain it for the majority of the week (to rule out games which trend for all of an hour, which isn’t much of a trend). The reason for trend is the most manual part of this infographic as I have to hunt down just exactly caused them to trend which can be rather esoteric on occasion.
The last section is a couple graphs of data that vary substanitally week to week and can speak to how the week’s games are fairing with the wider community. The hours spent playing, for example, shows the relative percentage of time players spent playing. The first section will always be the largest however the sizes of the sections change week to week. For instance I’ve seen the first section in one week span from 0 to 34 hours played whilst others only spanning 0 to 5. Weeks with broader sections would indicate that the games released that week were capturing players for longer. Similarly the last graph gives an indication of which genres were most popular, in terms of play time, during that week.
So, from now on, I’ll be providing these infographics once per week. I’m keen to see what you like about them, what extra information you’d like to see and what changes you’d make to them. If you’ve got any questions or comments feel free to hit me up on Twitter or through my public email address email@example.com.