This article explores the hypothesis that a user’s attention to a particular blockchain project affects its market capitalization deviation
Given high volatility of the crypto market, identifying the main factors of the blockchain project’s development becomes especially urgent to predict the market behavior. The influence of speculation in the crypto market prevents us from applying traditional research tools that are used, for example, to generate predictions on the stock market. A situation like this provokes the search for new solutions and metrics to identify crypto market trends. To solve the issue, we tested the hypothesis that a user’s attention to a particular blockchain project affects its market capitalization deviation.
The goal of the research is to identify how the specific crypto currency’s market cap deviation dynamics correlates with the Google Trends dynamics and the change in the number of Twitter followers.
As a reference variable, we used market capitalization deviation of each cryptocurrency calculated as
To calculate the number of requests for each cryptocurrency in Google, we used the public Google Trends data.
To measure the attention of Twitter users to particular cryptocurrency accounts, data on the Followers’ dynamics percentage was calculated as
For the current research, we used open historical data from Coinmarketcap.com that covers the period from 23 Aug 2017 to the end of February 2018 (market cap data). We also employed the data provided in the public domain Google Trends for the same period, as well as monitoring data on the number of followers of crypto projects’ Twitter accounts provided by monitoring services on social media profiles.
The graphs that we compiled help to track the dynamics of the indicated variables for six months for each of the blockchain projects under review (see charts below). We found no clear correlation between three considered variables. For some cryptocurrencies such as BCH, BTC, IOTA, LTC, their peak values roughly fall into the same time intervals, usually the end of November to December 2017. This is due to the global dynamics of the crypto market characterized by a sharp increase in the market capitalization at the end of 2017. For example, we observe roughly coincident growth periods for all three parameters for BCH, BTC, IOTA, LTC, QTUM, XLM, XRP.
For some cryptocurrencies, the cap deviation dynamics is observed in line with the market: it has an upward trend up to December with peak values in mid-December, then it declines either rapidly or gradually (BTC, IOTA, LTC).
Another group of cryptocurrencies (TRX, XEM, XLM) demonstrates the highest values of all three variables in the beginning – mid-January 2018.
Let us look at the correlation coefficients table (see below) to assess the strength of the correlation between the variables under consideration. For some cryptocurrencies, we observe a very weak correlation, for instance, XMR, VEN, LSK, QTUM. Another group of currencies is characterized by the weak correlations between cap deviation and two other variables, but rather strong correlation between two variables representing users’ attention (Twitter followers and Google Trends). This group includes BTC, ETH, DASH, ETC, XEM.
If we consider DASH example, there are several peak values for the cap deviation: at the end of August, September, mid-November, early and late December. Although, none of them correlate with the peak values for followers on Twitter. At the same time, the largest number of requests in Google coincides with the peak values of the cap deviation at the end of December.
The only case when high correlation coefficients were demonstrated for all variables is MIOTA. It is noteworthy that for some cryptocurrencies, there are negative correlations between the cap deviation and the variables representing users’ attention: ETH, ETC, NEO, VEN, LSK (for the last three – negative correlations are only between the cap deviation and the followers). This means that the growth of one variable is not related to the growth of another variable.
Overall, it is challenging to identify the leading indicators of the cap deviation dynamics because of the high volatility and speculative component of the crypto market. During the observation period, that covered the timeframe before the beginning of rapid market growth, peak values of capitalization and the subsequent decline, we found only a weak correlation between the market capitalization deviation and the interest of Twitter and Google users to the crypto projects of our research. In some cases, the dependence between variables related to a particular cryptocurrency is observed only in the peak period and is caused by the global noise around digital assets in late 2017 to early 2018.
To conclude, the problem of identifying the leading factors of crypto market`s dynamics remains open. In subsequent research, we plan to consider new variables (especially data on other social networks) that could predict the market capitalization deviation. We also plan to conduct regression analysis to identify the model with the greatest predictive ability.