Bitcoin Forum

Other => Meta => Topic started by: illiki23 on December 12, 2017, 12:57:58 AM



Title: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: illiki23 on December 12, 2017, 12:57:58 AM
This one creates larger word clouds.  Each consists of the texts for twenty posts.  Can you guess which one is more likely to be bounty hunters and which is the miner? ;)  

Candidate One
https://i.imgur.com/4V4FDTN.png

Candidate Two
https://i.imgur.com/NU16Bci.png


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: DarkStar_ on December 12, 2017, 01:03:14 AM
I'm going to say person #1 is the miner and person #2 is the bounty hunter /s


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: hilariousandco on December 12, 2017, 01:14:42 PM
Do one for me and my alt hilariousetc. I want to see if the word shitposters stands out like a sore cock haha (or anything else funny).


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: illiki23 on December 12, 2017, 07:08:51 PM
Do one for me and my alt hilariousetc. I want to see if the word shitposters stands out like a sore cock haha (or anything else funny).


Well here goes one of them.  I do see 'shitposters' in there, and oddly enough you don't actually say 'shitpost' or 'shitposter' but talk about them as a collective.  I put the word 'one' in the stopwords since you use it so much and it was so large, but it will be one way we can see that the two accounts have a similar writing style.  One neat thing is that your other account also says 'shitposters' but not the others as well and also says 'one' a lot.   There is a Kaggle contest going on right now where you train a classifier to guess who the author of snippets are using their writing for training. ;)

https://i.imgur.com/8UZDgMn.png

Currently refactoring that script which was put together in an hour of inspiration and hypomania.  It does not reflect thought out design. Will be able to show monthly trends after the refactor.  

Then I will compare your two accounts.


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: MadZ on December 12, 2017, 08:54:33 PM
Do one for me and my alt hilariousetc. I want to see if the word shitposters stands out like a sore cock haha (or anything else funny).


Well here goes one of them.  I do see 'shitposters' in there, and oddly enough you don't actually say 'shitpost' or 'shitposter' but talk about them as a collective.  I put the word 'one' in the stopwords since you use it so much and it was so large, but it will be one way we can see that the two accounts have a similar writing style.  One neat thing is that your other account also says 'shitposters' but not the others as well and also says 'one' a lot.   There is a Kaggle contest going on right now where you train a classifier to guess who the author of snippets are using their writing for training. ;)

https://i.imgur.com/8UZDgMn.png

Currently refactoring that script which was put together in an hour of inspiration and hypomania.  It does not reflect thought out design. Will be able to show monthly trends after the refactor.  

Then I will compare your two accounts.

I wonder who this "themos" person is  :D (top middle)


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: bluefirecorp_ on December 12, 2017, 08:59:18 PM
Can you publish your code on github or somewhere? It'd be interesting to see for sure.


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: illiki23 on December 12, 2017, 09:14:05 PM
Can you publish your code on github or somewhere? It'd be interesting to see for sure.

Sure.  I attached a script to a previous post but it was like written in an hour.  I am refactoring the code into small testable functions and adding some functionality so I can generate wordclouds based on data ranges and other things.  Once that is done today or tomorrow I will start a github for this and other data visualization projects and start breaking it out into modules.


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: bluefirecorp_ on December 12, 2017, 09:17:08 PM
Can you publish your code on github or somewhere? It'd be interesting to see for sure.

Sure.  I attached a script to a previous post but it was like written in an hour.  I am refactoring the code into small testable functions and adding some functionality so I can generate wordclouds based on data ranges and other things.  Once that is done today or tomorrow I will start a github for this and other data visualization projects and start breaking it out into modules.

Sounds amazing. Thanks for your hard work!


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: Lauda on December 12, 2017, 09:24:16 PM
Looks very interesting and I do tend to like visualizations. If you have the time, please do one for me as well.


Title: Re: Larger word clouds. Can we guess who is the bounty hunter and who is the miner?
Post by: illiki23 on December 13, 2017, 09:25:39 PM
Can you publish your code on github or somewhere? It'd be interesting to see for sure.

Sure.  I attached a script to a previous post but it was like written in an hour.  I am refactoring the code into small testable functions and adding some functionality so I can generate wordclouds based on data ranges and other things.  Once that is done today or tomorrow I will start a github for this and other data visualization projects and start breaking it out into modules.

Sounds amazing. Thanks for your hard work!


ANN thread visualization and analysis?  ICO analysis?
User Post History analysis?  Safely and Benignly Crawling Bitcointalk?
Shitpost modeling?

If any of these things interest you come join me in developing an Open Source project.  https://github.com/pandora23/bitcointalk-visualization/blob/master/README.md

https://trello.com/b/fBefa8LD

Slack: bctminingcollective.slack.com