Do more active coders leave longer or shorter comments than others?

Readme

This is a query that, when scatter-plotted, will show that more active coders
(as measured by those who have a higher number of pushes announced on the GitHub timeline)
do not necessarily leave much longer or shorter comments than others
(in my run, the correlation was positive and statistically significant: r=.166, df=15997, p<.0005,
statistically meaning that more active coders leave slightly longer comments),
but have lower variance in their comment length. 
The scatter plot looks sort of like Bullet Bill from Mario brothers. 

This is not the original query that I ran,
but modified to be a working query doing all the transformations
and data selection that I did.  

    

Query.sql

SELECT actor,
LOG10(COUNT(actor)) LogCount,
LOG10(AVG(LENGTH(payload_commit_msg))) LogCommentSizeMean
FROM [githubarchive:github.timeline]
WHERE repository_private='false'
AND type ='PushEvent'
GROUP BY actor
LIMIT 16000;