Forecasting various economic indicators has been a primary interest in economics and has attracted the attention of many researchers. Granger causality analysis has become quite popular in the econometrics literature and it aims to determine whether one time series is useful in forecasting another. In this work through the use of Granger causality analysis we investigate whether Twitter sentiment, expressed in large scale collections of daily tweets, can be correlated or even predictive of future prices of cryptocurrencies. The proposed framework considers tweets that mention the cryptocurrency “Dogecoin” and analyses the textual content of each of these tweets using a modified version of the lexicon-based sentiment polarity analysis method, VADER. The generated, Twitter sentiment time series is then compared to a time series of the closing prices of Dogecoin for each day. Granger causality analysis showed a unidirectional relationship between Twitter sentiment and cryptocurrency prices for day lags ranging from 2 to 4 days (with a 3-day lag having the lowest statistical significance value). This was also accompanied by a Pearson correlation coefficient of r= 0.6940 and a clear visual correlation between the two time series (with this 3-day lag). Findings indicate that Twitter sentiment is directly correlated and can be predictive of the future prices of cryptocurrencies.