1082651 - Question 5: Can we create a geographical overview of our contributors. i.e. by country.

Status: NEW

(Community Building :: Systems and Data, task)

Description

[Adam Lofting (:adamlofting)]

Moving this question from Rina in the wiki into a bug so it doesn't get lost.

"Can we create a geographical overview of our contributors. i.e. by country. "

Comment 1

[Rabimba]

Do we have a (unified)data source?

(In reply to Adam Lofting (:adamlofting) from comment #0)
> Moving this question from Rina in the wiki into a bug so it doesn't get lost.
> 
> "Can we create a geographical overview of our contributors. i.e. by country.
> "

Comment 2

[Adam Lofting (:adamlofting)]

This ticket is part of a joint MoCo/MoFo contributor analysis project. Find out more here: https://wiki.mozilla.org/Contribute/analysis 

We are using one ticket per question to track this work. Our goal is to answer a number of questions before the co-incident workweek. Some questions will not be practical to answer in this timeframe and when this happens we will keep the tickets for ongoing analysis after the workweek.

If you would like to work on this question, please comment on the ticket.

---

Comment 3

[Adam Lofting (:adamlofting)]

Some notes about potential data sources:

* We have self-reporting for core contributors in Mozillians.org
* We have event location for webmaker mentors who host events 
* We have Reps data

Comment 4

[rina]

Hej Rabimba 


Unfortunately not. Reps homepage, FSA homepage are some good places to look. I will add TJ to this bug as well as he has some of the stats by country.

By country we have 
Reps
FSA
Social media (FSA/reps/launch teams)

@adam, is there other groups to consider? 



(In reply to Rabimba from comment #1)
> Do we have a (unified)data source?
> 
> (In reply to Adam Lofting (:adamlofting) from comment #0)
> > Moving this question from Rina in the wiki into a bug so it doesn't get lost.
> > 
> > "Can we create a geographical overview of our contributors. i.e. by country.
> > "

Comment 5

[Adam Lofting (:adamlofting)]

There is lots of general data about geographic activity (for users and supporters) - e.g. anonymous web analytics tell you how many visitors come from each country.

We do have a unified data source for contribution activity (https://wiki.mozilla.org/Baloo), but when you get down to an 'active contributor' level of engagement, most of those systems don't have geo-data (and quite rightly so).

So, for example, we don't know where in the world someone is when they commit a piece of code, or answer a support question on SUMO. 

- 

I don't know what data FSA collects but we should investigate as part of this task:
https://www.mozilla.org/en-US/contribute/studentambassadors/

Comment 6

[Rabimba]

(In reply to Adam Lofting (:adamlofting) from comment #3)
> Some notes about potential data sources:
> 
> * We have self-reporting for core contributors in Mozillians.org
> * We have event location for webmaker mentors who host events 
> * We have Reps data

That should be pretty good indicator for now. Assuming we can somehow co-relate a webmaker mentor, rep and a mozillians profile on some common vector, like irc nick/email/name (or combination). In that case we can take geo location from whichever profile has it.

(In reply to rina from comment #4)
> By country we have 
> Reps
> FSA
> Social media (FSA/reps/launch teams)
> 
If we can co-relate data/profiles between them then we should be able to get a pretty good idea.

(In reply to Adam Lofting (:adamlofting) from comment #5)
> So, for example, we don't know where in the world someone is when they
> commit a piece of code, or answer a support question on SUMO. 

If we can do that above co-relation then we should be able to categorize commit/support question answer on SUMO by location, right?
Also we should be able to get pretty good idea about active and dormant contributors (active = code commmit/sumo activity+mozillian account, dormant = only mozillian account but no other activity)

Comment 7

[Rabimba]

Also on a side note we should be able to relate reps events with contribution activity.
Just an idea, to track down the impact of various reps events with contribution impacts. Maynot be very accurate, but with reps portal event location and comparing it with active contributors(active = code commmit/sumo activity+mozillian account) joining in next couple of days (maybe 7 days or 14?).

We should be able to loosely track impact of events.

Comment 8

[Adam Lofting (:adamlofting)]

We're not going to be able to join up these data-sets to see how much they overlap is in the next few weeks, but we have 3 sources of data we can compare and try to look for clues in.

1) Reps by country:
https://docs.google.com/a/mozillafoundation.org/spreadsheets/d/1bDd1_GZDCKLespa_NnJoG-bLjVKwo74VxKIlWG0xj94/edit#gid=0

2) Webmaker Event Hosts by country
https://docs.google.com/a/mozillafoundation.org/spreadsheets/d/1bJNUoowACvUNNScLNnEwRUWULm5zXI1E98AbSU_YCwY/edit#gid=0

3) Mozillians.org self-reported county 
I will file a new bug to extract this

Some things that might be interesting is to normalize these numbers against country populations (or even population with access to the internet) For example: http://en.wikipedia.org/wiki/List_of_countries_by_number_of_Internet_users

@Rabimba, would you like to look at this data? (anyone else is welcome too).

Feel free to make copies of those Google Docs and edit or work on the data any way you like.

Comment 9

[Rabimba]

Yeah I will be happy to play with these. 
Just the immediate output form them plots up to something like this (both are just for reps right now)

http://imgur.com/fZjf54r
http://imgur.com/xpNxcDZ 

I will play with them a little more with the idea you gave (population of country as well as with access to internet, we can get penetration that way).

Also I was thinking can I get a little more data dump? If that can be exported?
What I have in mind is maybe a sheet with the following rows/columns

Name || Country || email/irc (anyone of them, your choice)

In that case I will be handle the overlap at-least to some extent and the visualizations can be much more fine grained and interactive (we should be literally able to dive down).

And it will be very interesting to see how the data develops/changes over time.

(In reply to Adam Lofting (:adamlofting) from comment #8)
> 1) Reps by country:
> https://docs.google.com/a/mozillafoundation.org/spreadsheets/d/
> 1bDd1_GZDCKLespa_NnJoG-bLjVKwo74VxKIlWG0xj94/edit#gid=0
> 
> 2) Webmaker Event Hosts by country
> https://docs.google.com/a/mozillafoundation.org/spreadsheets/d/
> 1bJNUoowACvUNNScLNnEwRUWULm5zXI1E98AbSU_YCwY/edit#gid=0
> 
> 3) Mozillians.org self-reported county 
> I will file a new bug to extract this
> 
> Some things that might be interesting is to normalize these numbers against
> country populations (or even population with access to the internet) For
> example:
> http://en.wikipedia.org/wiki/List_of_countries_by_number_of_Internet_users
> 
> @Rabimba, would you like to look at this data? (anyone else is welcome too).
> 
> Feel free to make copies of those Google Docs and edit or work on the data
> any way you like.