@paulag what I can said is this is a very good example of selective presentation of data to frame people. Nice work!
First of all, I have to declare my interests, I am one of the guy who @htliao, @nicolemoker and @linuslee0216 often voted, this was because we are from Hong Kong and we joined Steemit at almost the same time due to introduction of same person.
I would like to say you've dig out a lot of data which is the value of this post. However, it is clear to me that your conclusion had been deliberately set to make other people think to the direction you would like to, which is a very commonly used technique in business world to "achieve your goal by just selectively present data favour your conclusion". Again, this was nicely done and I believe many people who didn't care about the turth would very much like to see your conclusion.
Let's see that "Fact" that you use to conclude @htliao was selling votes, it was messages from @aafrin sent to @htliao asking for upvotes, sounds good proof! But let's see what @aafrin did:

Our fellow steemian like to send lots of messages to others who got plenty of SP for upvote. So I guess it is very natural for @aafrin to do the same thing to @htliao.
Guess what? I think he didn't get anything he needed in return that's why for later 18 days, no one went to @htliao for upvote business! What a failure @htliao was in the selling votes business!!!!! Maybe @ned undelegated his SP because @htliao could not sell votes well?

Let me show you all something else, there was another Steemian did the same thing by sending 1SBD to @htliao and ask for vote, and @htliao didn't do anything!!!!

What a bad businessman @htliao was if he was selling vote! I think you did the right thing by point out @htliao was not doing good in selling vote! No wonder @ned took the SP back!
I think @ned likes people who know how to use bot to vote and for those who really care about contents by looking into the posts before they votes, which made them cannot votes over thousands time per day, @ned thought them are not effective and took the SP back! That's the concluson I draw from your data. Thanks for the data!