Using social media signals to invest in cryptocurrencies -> First results

[cryptofan2017]

A while ago I created a social media scorecard which scores coins based on their social media sentiment (Twitter, Google, Reddit and Github). So far the results of this are pretty good (about +134% first week for top 3 coins, 113% second week). Though in my own opinion scorecard should be used as a support signal, these are interesting results. Maybe somewhat superfluous but im completely aware that this sample size is way to small to make significant claims, and that these results do not guarantuee anything for the future.

Basis for starting this project are these papers

Algorithmic Trading of Cryptocurrency Based on Twitter Sentiment Analysis
http://cs229.stanford.edu/proj2015/029_report.pdf
This paper predicted the hour to hour sign change of the bitcoin price using Logistic Regressions, Naive Bayes and SVM.

Predicting Fluctuations in Cryptocurrency Transactions Based on User Comments and Replies
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4988639/

Im curious as to what your experience is on trading on social media? Do you use it and in what way?

In case one is interested in the full report I use, it is available at www.AltcoinAnalytics.com

[saidlik]

 Very nice, i think there should be some really nice results

[cryptofan2017]

Thanks, first results are indeed promising:)

[asdalani]

A while ago I created a social media scorecard which scores coins based on their social media sentiment (Twitter, Google, Reddit and Github). So far the results of this are pretty good (about +134% first week for top 3 coins, 113% second week). Though in my own opinion scorecard should be used as a support signal, these are interesting results. Maybe somewhat superfluous but im completely aware that this sample size is way to small to make significant claims, and that these results do not guarantuee anything for the future.

Basis for starting this project are these papers

Algorithmic Trading of Cryptocurrency Based on Twitter Sentiment Analysis
http://cs229.stanford.edu/proj2015/029_report.pdf
This paper predicted the hour to hour sign change of the bitcoin price using Logistic Regressions, Naive Bayes and SVM.

Predicting Fluctuations in Cryptocurrency Transactions Based on User Comments and Replies
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4988639/

Im curious as to what your experience is on trading on social media? Do you use it and in what way?

In case one is interested in the full report I use, it is available at www.AltcoinAnalytics.com

You could also look at how the coins are moving in coinmarketcap. There is a lot of research that is needed before investing so you might not want to stop there.

[cryptofan2017]

Actually the analytics incorporate market caps. The logic scales social media exposure relative to their market caps. Have a look at AltcoinAnalytics.com to see total logic of the report.