Thank you to everyone for their feedback and initial impressions on our software, it really does help us out a lot so thank you.

I'd like to address a couple of points many of you brought up just to perhaps clarify some doubts, since we have been thinking about this project for some time now, and have developed strategies to circumvent some of the stuff you guys mention here.

The most common concern seems to be that it is very difficult to differentiate from spam, artificial hype, fake pump schemes etc. This is all true. However, the way we track social media mentions and sentiment helps us remove most of the spam accounts, bots, etc. In the case of twitter for example, we have many different rules to determine whether we should count that tweet as "real". These exclude tweets that have account follows less than 5, frequent spam reports (twitter notifies us), bot accounts by twitter mandatorily need to have a specific tag in the bio, so we filter out tweets with it, if they follow many reported spam accounts, if they repeatedly copy and paste messages etc. All of these rules add up to getting the most accurate representation online.

It's worth mentioning that the enterprise API twitter has is very robust and they have very thorough guidelines on how to get the most accurate representations. If it was full of spam, it would be pretty much useless, and considering how expensive it is, it wouldn't be good business for them.

The second most common concern is that tracking sentiment isn't a good strategy for investment. While I agree that for Bitcoin and some of the bigger market-cap coins this could be true, for many of the alt-coins where young people like to join in a community and collectively ascribe it value, it's a very useful indicator of how things evolve in the short term. Once we had a working prototype for our software, we ran some correlation tests and price movements from coins on our trending page most of the time correlated with price movements later on in the day with varying degrees of lag. The studies are clear on this (I'll link the ones we read before starting this project below), social media sentiment has a very high correlation with price movements in areas such as the stock market. Of course, this whole project is an experiment on whether the same is true for cryptocurrencies (which IMO are predicated on mostly market sentiment or "hype") but so far, the correlation seems to stand and is quite strong. A lot will improve in the future when we refine our AI more and provide even more accurate representation.

On another note a couple of you mentioned the design looks suspicious. I agree it's not the best, but so far we wanted to provide functionality over anything else. Everything will improve substantially in the future.

I hope I was able to inspire some confidence in our platform, as we believe it could be useful for everyone. We're not saying invest solely on this, but keep it as a valuable tool in your arsenal.

Thanks again everyone, Povilav

Hey guys, I want to let you know about a new software tracking market sentiment for cryptocurrencies called GrapevineCrypto.  It tracks social media market sentiment for all major coins (Bitcoin, Ethereum,....) and hundreds of smaller coins (approx 1000) in total.  As social media hype influences prices quite drastically with our software you can track how market sentiment changes and evolves, as well as keep an eye out for smaller altcoins that are on the rise.  If this interests you please check us out at:  https://www.grapevinecrypto.io , if you have any feedback please either reply to this post of fill in the form on the website.  Thank you all for your time and I hope you all have a great day.


EDIT

Thank you to everyone for their feedback and initial impressions on our software, it really does help us out a lot so thank you.

I'd like to address a couple of points many of you brought up just to perhaps clarify some doubts, since we have been thinking about this project for some time now, and have developed strategies to circumvent some of the stuff you guys mention here.

The most common concern seems to be that it is very difficult to differentiate from spam, artificial hype, fake pump schemes etc. This is all true. However, the way we track social media mentions and sentiment helps us remove most of the spam accounts, bots, etc. In the case of twitter for example, we have many different rules to determine whether we should count that tweet as "real". These exclude tweets that have account follows less than 5, frequent spam reports (twitter notifies us), bot accounts by twitter mandatorily need to have a specific tag in the bio, so we filter out tweets with it, if they follow many reported spam accounts, if they repeatedly copy and paste messages etc. All of these rules add up to getting the most accurate representation online.

It's worth mentioning that the enterprise API twitter has is very robust and they have very thorough guidelines on how to get the most accurate representations. If it was full of spam, it would be pretty much useless, and considering how expensive it is, it wouldn't be good business for them.

The second most common concern is that tracking sentiment isn't a good strategy for investment. While I agree that for Bitcoin and some of the bigger market-cap coins this could be true, for many of the alt-coins where young people like to join in a community and collectively ascribe it value, it's a very useful indicator of how things evolve in the short term. Once we had a working prototype for our software, we ran some correlation tests and price movements from coins on our trending page most of the time correlated with price movements later on in the day with varying degrees of lag. The studies are clear on this (I'll link the ones we read before starting this project below), social media sentiment has a very high correlation with price movements in areas such as the stock market. Of course, this whole project is an experiment on whether the same is true for cryptocurrencies (which IMO are predicated on mostly market sentiment or "hype") but so far, the correlation seems to stand and is quite strong. A lot will improve in the future when we refine our AI more and provide even more accurate representation.

On another note a couple of you mentioned the design looks suspicious. I agree it's not the best, but so far we wanted to provide functionality over anything else. Everything will improve substantially in the future.

I hope I was able to inspire some confidence in our platform, as we believe it could be useful for everyone. We're not saying invest solely on this, but keep it as a valuable tool in your arsenal.

Thanks again everyone, Povilov