Topic: Is the guy running the Bitcointalk Visualization Project absolutely insane? (Read 247 times)

Not everything is about profit my potential friend.  Some things are about taking one step at a time and living life on life's terms.  (sorry just been saturated with AA sayings lately)
Yeah eventually I will get filthy rich (ugh would have been by now if I didn't sell my tokens) it is just I want to have fun in the process.  And I guarantee if given the opportunity I can do something new and unexpected.

In the meantime is there anything I can do to entertain you within the context of visualizing bitcointalk forum threads?  (lots of cools stuff...)

Yup it is Unity.  As I recently decided to learn Unity.  And what you are seeing is the result of one day of programming as I need to rebuild everything in C# (I am used to Beautiful Soup and Splinter... ).  The Bitcointalk Visualization Project was something I created a few years back to post about and earn my signature credits as I was in grad school doing research on NLP (incongruity modeling/humor theory) involving visual data mining (visualizing the data in such a way we can identify model features) but like two weeks in someone completely unrelated to the forum offered me a job and I just kind of dropped the idea before getting far in.  Since then and a couple jobs later I decided to get into game development but well want to do that too...  

I was tempted to just call it the Bitcointalk Data Mining (or Analytics) Project or something but well visualization and visual data mining is fun so I was going to keep it to visual data mining. (Specifically 'visual text mining'...)  And really 'visual text mining' does use a lot of traditional machine learning approaches for feature generation which means I might engage in and discuss all sorts of traditional analytics approaches.

Here is a list of classification and prediction tasks we might want to explore: shilling, shitposting (also the influence of signature campaigning on post quality), bot detection, bitcointalk specific topic modeling and detection, and well just various visualizations related to sentiment and semantics (keyword/ngram features) that could be useful for any number of things or just tell a neat story???  

To be honest I actually have another project and vision but I would rather take this stupid opportunity to learn game development and get all my bad coding practices out of the way before starting that one..  It is a VR game.  And who knows where it will take me (hail eris please be kind).

Anyhoo, yesterday I started the 'stopword remover' but didn't have enough to post about.  It is basically an old soviet washing machine and when you drop the stack of 'users posts' into it is generates a new stack with stopwords removed.  And of course it literally spits the stopwords out into a pile.  Hey, who doesn't want to dig through stopwords?

Though Cosmos is holding a hackathon and one of the areas is game development using Strange Clan's NFT system and it is soooooo tempting.  But yeah I have wanted to finish this project for a long time and since starting literally taught data analytics so have a whole different perspective.  And it is just such a fun topic...  

Um... fun fact... I used to be into... neuro.... linguistic... programming.......  but yeah we are talking about natural language processing....

Is this Unity?

I don't understand what you were trying to achieve there. What's the visualization project exactly and why have you stopped working on it? I only see a table and texts hovering into the game engine's space.

Exactly what a machine learner plus neuro-linguistic programmer plus data analyzer would say. Sorry, I could not hold back.

If it's related to bitcoin by anyway, you can post it in that board. The importance of choosing the right board isn't just to make things well organised, it's also benefecial to your project itself as it's published in the board where users who hang out there may be interested to it.
As you can see, your topic is still in Meta and Mods don't think that it's necessary to move it. You can always move it manually (the mov button is in the bottom of this page). Good luck

Happy to put it wherever... just the first post on "Project Development" says it is SPECIFICALLY about Bitcoin related projects and not altcoins.  And this just doesn't fit that criteria.  It is about both as it is about anything related to bitcointalk analytics.  The best fit really is meta as it is about mining bitcointalk threads to perform analytics of our users.  I am more than happy to move this to another board just trying to figure out what.

Even if you do it for your own fun, this is a high time consuming project. Would it be profitable by anyway?

Then better to move it to "Project Developement" sub-board. This is not a Meta topic.

Not yet though there might eventually be...  I do like generative adversarial networks (and other deep learning approaches) but I am starting with decision/regression tree ensembles as they tend to be better at telling stories (easier to interpret).  Though some visualization approaches can use any type of classifier so eventually want to find the approach with the best accuracy and such.  What would you want to see done with these?  Any classification or predictions tasks in general anyone would like to see?  The sky is the limit (um literally the way I am doing this) and I am willing to go in directions based on demand. 

There are so many possible data analytics tasks.  To me I like ones which help people make wise decisions when it comes to investing and getting bamboozled.  But I also am going to go for some ones just for the purely storytelling aspect. 

The whole 'doing it in Unity VR whatever' thing is going to stretch this out quite a bit... but that is ok...  so I am probably going to be posting on NLP basics for awhile as I literally build a pipeline... Things like stopword removal, hyponym replacement, ngrams, tf-idf or word2vec word frequency associations, basic model training... oh and a word cloud generator..

Oh, I thought you were feeding the individual user data into a text based generative adversarial network or something.

Um I am kind of doing this for fun and as a way of analyzing Bitcointalk forum threads in an attempt to do some personal NLP based academic research by modeling a few phenomena relevant to this forum.  I am not providing a service to anyone.  I do want to share the results and definitely will take requests and yeah I will make everything open source.

I was actually going to ask where this awesome nonsense ("bullshit makes the flowers grow") should go.  It is not 'bitcointalk forum software', it is not a service, it is not off topic.  The best guess I had was meta but am open to suggestions.


I don't understand the question.  Right now I am just building the post crawler in a silly but well workable manner, there will be a user interface to input the parameters for the crawl.  I literally started yesterday so have to start somewhere.

I was going to start by something fun, just to get some pieces in place, so word clouds of various coins or different types of users like miners versus ecenomic traders.

But where it gets more interesting is doing more sophisticated modeling so that the bitcointalk forum data can tell stories I can then share with you all.  For example we could take and build decision tree ensembles (great for heterogeneous feature sets) which classify user types or do topic modeling.  And then of course literally invert them and apply leave and bark textures (you will still be able to read off the splitting criteria) because I am learning game development now...

Edit:  Bitcoin projects also doesn't work as I am probably going to focus on ANN threads... 

What is this project's significance to the bitcointalk forum? Shouldn't this be on the service sub-board, since it sounds like you're offering a service?

Edit: Thanks for your explanation the project is interesting keep it up.

Train the network with just my posts and gib a paragraph?

Well you are going to have to read my threads to find out..  

Anyhoo I have decided, as a way of posting in an informative and entertaining way, to restart the Bitcointalk Visualization Project but there was a dilemma - I had just recently decided to go into game development and want to take every opportunity to learn that.  That was when this ridiculous idea began, to do the entire BVP which involves some serious ML, NLP, and data analytics within some weird and ever evolving VR world.  And not going to leave this thread started with just that, here are my first screenshots:




As you can see I started tinkering on a prototype 'user post history' crawler on my workbench.  I think it would be funny if I had to prod it a couple times to wake it up and get it going.  Though the idea of having to shovel away discarded stopwords from the stopword remover....

As for next steps well there are just so many ways to go.  Feel free to leave input.  In terms of overall large features I got to get the word cloud generator up because it would be cool to well actually make word clouds.  
I do want to set up a hall of ANN thread word clouds for myself, specifically for ones running bounty campaigns, as I would like to see if there are any new interesting ones with a glance.