During the development of statscloud, we've hit a few big milestones along the way; our in-built R terminal, our shiny Dashboard view, the rollout of PRO accounts, and today we've hit yet another: You can now perform Bayesian analyses in the app!

🎉 Whoop!

Now, this isn't to say that you should be running Bayesian analyses on your projects. Of course not. It's entirely up to you whether to use frequentist or Bayesian analyses. No-one working on statscloud has any ulterior motives here; we didn't create the app to push you towards any statistical philosophy in particular because that would be wrong.

In the spirit of being honest, though, I feel I should come clean: I've never really been a fan of Bayesian analyses. I'm sorry. It's not that I disagree with the approach or think that they don't provide a useful way of analysing your data, it's actually for a much more simple reason:

They're really, really complicated.

I tried teaching myself the Bayesian approach several times over the past five years, and it always resulted in the same outcome: crawling under my desk and sobbing. Honestly, I just couldn't get my head around it. You may as well have tried teaching it to me in Portuguese.

However, there's no getting around the fact that Bayesian statistics are becoming more and more common now, so there was no way we could avoid having them in statscloud. Like it or lump it, we needed to have them in the app. So, when tasked with adding Bayesian tests in statscloud, I asked myself one question - the exact same question I've asked myself since day one of creating the app - "what tools do I need to help me understand this"?

As I've said many times before, the problem with statistics software today is that it assumes everyone is a genius. Each analysis has a million tick boxes you need to fight your way though just to run a simple test and, just when you think you've cracked it, you open up a dropdown box and then find a million more. Selecting one of these options can do anything from adding a star to a number to adding in an extra twenty tables, and it's up to you to trawl through the output and see what extra thing it spat out.

This is what separates statscloud from other statistics packages. It doesn't matter if you're a statistics genius, an amateur who needs a refresher, or someone entirely new to the field. You can still run an analysis following our simple wizard and make sense of it as you read through the output. Not only this, it's designed to teach you as you go along, it doesn't overwhelm you buckets of tick-boxes, and doesn't make you feel utterly helpless.

When you run a Bayesian test in statscloud, you'll see exactly the same interface you're used to when running any other test. You select the variables you want to test, hit "Run Analysis", and see everything summarised for you in an easy-to-understand format. If it's the first time you've seen a prior-posterior curve, then don't worry; the output tells you exactly what it is and how to interpret it. If you're wondering what a "Sequential Analysis" is, just click on it and the app will tell you in a nice digestible way.

This is exactly what I needed.

And we haven't even talked about the best thing about all this: with statscloud, you can run Bayesian analyses on any device at any time for free. This follows the same basic principle I've wanted to employ in statscloud from the start; to make stats free and accessible to everyone at any time. We haven't hidden these tests behind a paywall, and you don't need a fancy computer to run them.

At the moment, we've just released some basic Bayesian analyses (e.g. t-tests), but we'll be rolling out some more advanced ones in the future. For now, statscloud provides a really nice way of introducing Bayesian analyses to newcomers and will slot in really neatly to your existing statistics training.

I'm really proud of what we've created.