On Using LinkedIn to Analyze the Size of the Drupal Community

The Drop Times was started with a vision of contributing to the growth of the Drupal community. There are two aspects to this - one is focusing our efforts towards contributing towards the growth and the second is measuring the extent of the Drupal community globally and seeing how the community is doing numerically.

As a first step to working towards helping the global community grow we wanted to get a numerical sense of the spread of the Drupal community. However, the biggest challenge was to figure out where to get objective data on the spread of users.

Why not Drupal.org data?

Initially, we thought of using Drupal.org user data but there are three issues with that;

1. It is not public 
2. There are a lot of spam users registered on the site 
3. We don’t have country data for a lot of users. 

We wanted something we could quickly go back and check on later to see if the needle was moving.

Where else can we get data? LinkedIn?

The next option in front of us was LinkedIn. LinkedIn people search is public (for those with LinkedIn accounts) and it will allow you to search for keywords and filter by geography. But is this data accurate?

So the solution we came up with was to just use regular LinkedIn search and then search for the keyword “Drupal” and further filter by country (or city or state as needed) to get the number of users on LinkedIn who have Drupal as a keyword in their profile.

Whenever a “Metro Area” option is available in the LinkedIn filter for a given city we would use the Metro Area so that we include the users in the metropolitan area which would be the city and the suburbs. This is more meaningful when we try to use this data for organizing events etc...

Is this accurate? It is not, but it is alright!

Let us look at each of the problems with this approach and evaluate.

These people may no longer be working with Drupal

The search result will give you people who have Drupal in their profile. This does not mean that they are working with Drupal still.

This is alright because this error will be independent of the user’s location and hence the error should evenly be spread across all the locations. This would mean that we will still be able to use this number as a good measure for evaluating the spread and the growth of the community. The absolute values of these numbers will have errors but the number itself should be a good indicator.

Not everybody who is working with Drupal will be on LinkedIn.

There are supposedly more than a billion members on LinkedIn and it is reasonably well spread throughout the world and more so across technology users in the world. This is definitely a sizable part of the employed/employable population. It very likely covers the vast majority of technology users across the world. It still will not include every technology user in the world.

As with the first error, this error will also be independent of the user’s location and hence the numbers from these searches can continue to work as good measures/indicators for the spread and the growth of the Drupal community.

Not everybody who works with Drupal and who is on LinkedIn will have Drupal in their profile.

This is also true and there will be errors associated with this. As with the earlier errors, this error will also be independent of the location and the metric will continue to be a good measure.

Just having Drupal in the profile need not mean that they are involved/interested in the Drupal community

We wanted a simple measure that can be re-evaluated quickly when we check for the growth/health of the community later or when we want to decide where to focus our efforts.

We thought of using combined keywords like Drupal Developer, Drupal Architect, etc... but we will never cover all the relevant variations and it would make the search very cumbersome to run and re-run.

Numbers above 1000 are rounded off to the nearest 100

LinkedIn only shows the rounded-off numbers for search results greater than 1000. So there is a 2-10% error on these numbers above 1000. This is an error we will have to live with and this affects 40-odd cities.  

The numbers themselves are meaningless

This is also probably true in an absolute sense because of the above errors. However, every one of these people who have the keyword Drupal in their profile knows about Drupal. These numbers may have errors in an absolute sense, but they become meaningful when compared across different regions and over time.

For example - the percentage of global Drupal users in a geographical area will be reasonably accurate, the relative distribution of users across the world will be reasonably accurate, growth of Drupal users in a given location across time will be reasonably accurate.

This would make these numbers very useful from the perspective of decision-making when we try to organize local communities or events and from the perspective of evaluating the growth of the community through our activities later.

Our analysis of LinkedIn data reveals that 81 cities globally each have over 500 people with "Drupal" in their profiles, indicating strong local Drupal communities. Please read, DrupalCollab: Drupal Community in the Largest 500 Cities in the World.

Are there other issues with this approach?

We can’t think of any other issue with this approach but if you do please reach out and let us know and we can try to address those.