How Many Google+ Users So Far?

I have completed an initial analysis of the current size of the Google+ userbase using surname and population data from US Census Bureau. I will explain later how my model works and how it is flawed, but here is what I have learned so far:

With 6 popular surnames in the model, the estimated Google+ user base was 645,000 users.

With 39 surnames, the model estimated 593,000 users.

And with 50 surnames, the model estimated 587,000 users.

The model's output doesn't change wildly as more surnames are analyzed. That is reassuring. However, my model is flawed. I'm using US surname popularity data, finding how many Google+ users there are with each surname, and then extrapolating to total users worldwide. I tried to limit the surnames I analyzed to those that are highly concentrated in the US, but I need to try harder. I included "Kaufmann" for example, which gets a lot of matches from Germany that I didn't omit. Using that surname alone the model shows 2.1 million G+ users worldwide. Thus, my model overestimates worldwide users when it uses surnames that are not only common in the US but common elsewhere.

In brief, here is my model:

The surname Cooke is the #996 most popular surname in the US. The Census Bureau estimates that there are 29844 Cookes in the US. I found 74 Cookes on Google+ so far. That would be .002479% of the total Cookes in the US. Taking the official current US population estimate of 311,677,540 and dividing multiplying it (thanks, Gaspar) by .00002479, you get 772,823 Google+ users, if all other surnames had the same distribution in the US as they do on Google+.

To improve the model, I need to either 1) find and use worldwide surname popularity data - instead of relying on US data only - or 2) I need to limit my surname matches in the Google+ user base to US only users, which would take a great deal more time. It would require looking at each profile page if the location was not apparent in the Google+ search results. And then, I would need to come up with a way to estimate world users of G+ to supplement my highly accurate estimate of US G+ users.

I'l probably use the latter approach for my next analysis, and I'll end up with an extremely accurate US G+ users estimate.

For the time being, given that my model overstates worldwide users because I have a number of surnames that are somewhat common elsewhere, I'm just going to irresponsibly guess that my model is flawed by 30% and so the number of Google+ users worldwide is exactly 410773.3, not including cats. :)

Before I blog about all this, I'll improve my model and have a better approach to US vs. Worldwide users.

But I think I'm done for today.
Shared publiclyView activity