How many web developers are there? What are the demographics?

[foolip]

Many discussions about web developer research or surveys raise questions around representation and bias. To help answer questions like this, it would be great to have a good idea about what the global population of web developers looks like, as a point of comparison. (This is not to say all research needs to strive to target the global population.)

Here's what I found when looking into this in October 2020:

World:

• ~16.3M in Q2 2019 (Apparently from SlashData's "Web developer population forecast 2020", but I have no link)
• lower bound 16.9M (the global developer population 2019 p. 15)
• upper bound 18.9M (the global developer population 2019 p. 8)
• So something like 0.25% of the world population (7.7 billion)
• Growing overall and changing with the economy, so we should really be thinking about this as a confidence “corridor” over time.

Reachable by surveys:

• lower bound 39k (Stack Overflow 2020 JavaScript users) which could be as little as 0.2% of developers

Europe:

• Eurostat should have data, SlashData refers to it.

USA:

• lower bound 103k (Data USA) or 174k (U.S. Bureau of Labor Statistics)
• upper bound 1.36M (Data USA: all software developers)
• SlashData must have a tighter bound

[foolip]

At that time I was mostly focusing on how many web developers there are. I think more important is to understand the basic demographics. Surveys (example) often ask about these some of these demographics:

• Country
• Language
• Age (often to exclude participants under 18 years)
• Gender identity
• Race & ethnicity
• Experience
• Salary

It would be useful gather and make accessible what we know (and don't know) about these. By gathering many examples, we could start to say what is an outlier and what is typical for some of these demographics. One step further would be to set confidence intervals for what the "true" numbers in the global developers audience is, which is probably not the same as the typical numbers in all the surveys. Specifically, I strongly suspect that men are overrepresented in most web developer surveys, but by how much?

[bkardell]

I brought this up in the meeting today, but I was thinking that it might be worth running a random survey on MDN - just a quick 1-3 questions that takes almost no time and see what it looks like.

The reason being that it is an extremely widely pointed at, very key and independent resource - the documentation of the commons. It's equally useful to novices and experts. So, it's not likely to have many of the other kinds of biases that have to do with companies, skill levels, who follows who or something geographic, and keeping it short should help avoid the "this is really a better measure of people who are willing to fill in 20 minute questionnaires" problem. Like, at a minimum it would be interesting to see how it correlates to other data.

In the meeting it was also mentioned that there was data from the MDN DNA survey from 2020