I think it's okay for us to know how many unshared projects there are, total. That's an interesting number, although what it means for a student project to be unshared is different for what it means for someone's non-school project to be unshared.
Yes, I guess we could look at unshared projects if we wanted to, but honestly, it's all we can do to keep up with the shared ones. :~)
In a perfect world, unshared projects would be stored encrypted. My colleagues keep explaining to me why we can't do that (in a way such that we couldn't decrypt them). I guess it would require you to store the same private key on all devices where you log in? How do secret message apps like Telegram work?
You have the same actualized graphic in this page https://scratch.mit.edu/statistics/.
It would be nice to have such graphics, not to compare with Scratch, but to know who is using Snap! all over the world.
It turns out I deeply offended Jens by talking about "lying with statistics." I've already apologized to him privately, but I should do it in public, too.
I did not mean to suggest that Jens was deliberately trying to mislead anyone about the strength and success of our project -- his project, mainly. Jens is a great person who has devoted his life to giving the gift of computer science education to young people.
What I should have said is that the statistics that are easiest to collect are rarely the ones that give rise to the greatest insight. This has been a huge issue for data science in general; in the recent development of a Data Science curriculum and degree program at Berkeley, my colleagues have built in a serious emphasis on data ethics, on how to find the best information you can from the ocean of data out there.
Right now, for example, the world needs statistics on how many people have died from COVID-19, which sounds at first glance like an easy thing to measure, but it isn't, because it turns out that how diagnoses are recorded on death certificates has different rules in different places. If someone with a preexisting problem such as diabetes is infected with covid-19 and then dies because of the failure of some organ that's affected by diabetes and also affected by covid-19, what's the cause of death? This is problematic not because anyone is trying to mislead, but because doing statistics is hard.
Our own, much less serious, data science problem has the same character. Nobody is trying to mislead, and specifically Jens isn't trying to mislead. I'm sorry if what I said made it seem otherwise.