The recent few years witnessed discovery of the predictive power of social media for various financial markets. Academic research has proven during the past few years that social media can provide investors across various financial markets with valuable information that aids them to formulate relatively accurate investment decisions.

Realization of the value of various social media networks led to the creation of novel methods to perform financial analysis through observation of the dynamics of social networks. For example, DataSift is a platform that is designed to collect and analyze unstructured data obtained from social networks such as Facebook. Also, Cayman Atlantic is an investment management startup that provide managed trading accounts that are based on real time data obtained from social networks such as Facebook, Twitter and others, and analyzing it via sentiment analysis.

A recently published study presented methods to differentiate between valuable information, regarding bitcoin price prediction, and noises through analysis of the effects of social network structures and their incentive hierarchy system. The study utilized data obtained from the bitcoin market, namely price data obtained from, and showed that the most visited social discussion networks are less accurate when it comes to prediction of future price movements, mainly due to information free riding, in addition to highly correlated information.

The study used as an example of a social network that specializes in bitcoin, and other cryptocurrencies, related discussions. Analysis of the discussions on the forum’s various boards, especially the “Speculation” board, led to some interesting results. The study found out that valuable information is more likely to be posted by users who start active discussions with other users of the forum. The study also investigated how social network incentive hierarchy systems can influence the motivation of users to share content; thus, affecting the information quality. Interestingly enough, the study showed that active users of, who hold high rank badges, are less likely to provide accurate predictions when compared to less active users, with lower rank badges. This can be explained be reduction in motivations of users after acquiring high rank badges, in addition to increased use of the forum for socialization purposes which increases the proportion of noise in their posts.

Tags: , , , ,

Leave a Reply

We use cookies to give you the best online experience. By agreeing you accept the use of cookies in accordance with our cookie policy.