Bitcoin Forum

Hey guys, I just wanted to share my website with you, it might be a tool to you as it is for me. Check it out at https://cointwitterscanner.azurewebsites.net (https://cointwitterscanner.azurewebsites.net). All feedback is welcome :)

Looks like a decent site with a basic design but very informative at the same time. I must ask how you are taking the data for these tweets whether based on hashtags or simply direct coin names being mentioned in tweets? None the less the statistics are graphs talk a lot about a particular altcoin over and its popularity over time and I'm impressed with your unique idea that you used here.

Hey thanks for the positive response 13abyknight, no, it's a search for {Coinname OR TickerSymbol}, so there is a lot of weird data hidden in there. I would not use it to find any meaningful data for coins as OMG and OKCash (OK) myself :) However, I think the website still has a lot of value in its current form. If you have any ideas how to refine the data, I'd love to hear. Hashtags would work, but I'm not sure whether each coin has it's coin name with spaces removed as a hashtag, since there are a lot of coins with long names. I might try that in the future, or do a simple search within quotes for the coin name. I'm doing regular updates on the website, so keep checking :)

I must say that the efforts you have given are quite remarkable. The idea is unique and hats off to that. The statistics you have given is neatly presented. The website looks very simple and could have been better. Try to design the website a bit and that will attract many users towards it.
The data given can be useful sometimes when one wants to see which coin is getting more hype on the social media and this can be useful for users whether to invest in a particular coin.

Very interesting tool, indeed!

I have heard about this concept before, using Trollbox messages to identify market sentiments. Of course, this can be applied to Twitter too.

The big question for me is, how valuable this tool is?

Looking at the graph for BTS, I can see that the Tweets for Oct. 05 and the price of BTS on Poloniex seems to be somewhat correlated. I can't help but wonder if the increase in Tweets for BTS is a result of the market movements or if it is the other way around.

Perhaps you could make an attempt to weight tweets, and maybe come up with a "score system" of sorts?

The Core team tweeting about an incoming update has far more potential significance than some shill twitter account with 80+% fake followers ;-)

Regards,

Very interesting. Though for now I couldn't think of any reason to use it, besides using it for giving me an idea on how Twitter-famous a certain cryptocurrency is. Could probably a decent indicator on how hyped up a certain coin is. Probably add a sort feature on the coins? A "top 10 this week" and a "top 10 in the last 24hrs" list?

Even though it looks to be sort of in the "beta" stage, I suggest working a bit more with the User Interface/UX. Trust me, that's a significant for your tool's success. Also, probably buy a cheap .info domain if you couldn't afford a .io or a .com domain.

Those are some good ideas, I'm working on it, and I have a few domains lined up (coinmarkettweet.com and coinmarketsocial.com) but they haven't been able to validate yet. The website is in deed not very mature, and needs a lot of things done. I'm working on getting some donation goals online now. I've been looking at your website, looking good :) A design shouldn't be too hard to implement for me, since I essentially only have two pages that need styling. Thanks for the tips anyway :)

Think of the word "awesome" - it generally correlates with positive sentiments (except in situations where awesome is used in its original meaning, then it's more neutral). Adding a "not" in front or behind the word "awesome" changes the sentiment of the phrase.
You can do sentiment analysis without using a NLP pipeline - there is no need to extract POS-tags, entities or dependencies if you have no need for them. In fact, you can just take a bunch of word vectors, and then average them over a sentence. Then softmax said average vector to your classes, and you'd get a very servicable sentiment analysis that would work 90% of the time. If you go down this path, I suggest using facebook's fasttext library which is..fast, and covers a lot of the challenges I mention below.

Try out using bootstrap. It's like the de-facto standard of user interfaces for years. It makes it alot easier too as you would only need to code minimal CSS, as most of the components you need are already in the bootstrap framework. Try it out! https://getbootstrap.com

My main advice for now is that you make your web tool less "boring". It just looks to generic. Also, provide some useful information like I said. Charts, graphs, pie charts, those stuff. :P

DISCLAIMER: I am in no way affiliated with bootstrap

This is some great info - i would see it as being especially useful when lookng for pump and dump schemes - what do the percentages indicate ? i assume some variance from a norm - but what norrm/average ? the last 15 minutes, the previous 15 minutes etc

It would be nice to be able to sort them by either positive or negative percentage

Craig

Yes this is very interesting to see, but i think you should work a little bit more on the design of your website, because it would be better if it gets more attractive, anyway, this is a very good tool for coin developers and those who want to see volumes from various coins.
I think that you will receive more replies and more attention if you move this thread to service disscussion.

Thank you for sharing. A interesting way to see the latest "hype". Will look to see if there is any link between tweets and price movement.