The impact of Twitter user sentiment on Bitcoin pricing value
Nguyen, Vy (2021)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2021120824603
https://urn.fi/URN:NBN:fi:amk-2021120824603
Tiivistelmä
Bitcoin is a cryptocurrency whose unique features have made it helpful in improving people’s lives for the better. Thanks to its decentralized nature, Bitcoin allows its users to transact and store wealth without the intervention of a third party that they may not trust. Despite its positive contribution to society, Bitcoin’s price is known to be heavily volatile. Regarding some of the causes behind its price fluctuations, the opinion expressed on social media of a highly influential figure, namely Elon Musk, seems to be one that fuels its volatility. Recognising this problem, the thesis aims to determine whether the opinion of the mass on social media can influence the price of Bitcoin too, or it is just the force of one influential individual.
The thesis uses a deductive research approach. Both theoretical framework and empirical research are developed using secondary data. In its theory part, the thesis utilizes previous research about Bitcoin price formation and key investors as well as social media’s impact on decision-making and investor behaviour. Based on the findings from these researches, the thesis draws a theoretical conclusion to whether social media sentiment can influence the behaviour of Bitcoin investors, which ultimately will transfer into changes in Bitcoin price. Using Twitter as a proxy for social media, the thesis performs VADER sentiment analysis for tweets about Bitcoin collected between July 18, 2021, and July 30, 2021. The thesis then performs the Granger-causality test and vector autoregression analysis to determine whether there is a statistically significant relationship between Twitter user sentiment and Bitcoin pricing value.
Based on the thesis’s empirical findings, it seems that there is not enough empirical evidence to conclude that Twitter sentiment has a statistically significant impact on Bitcoin price. Future studies are encouraged for two main reasons. First, data with a longer timeframe or shorter frequency might yield more statistically significant results. Second, there might be another social media platform that is a better proxy for social media than Twitter.
The thesis uses a deductive research approach. Both theoretical framework and empirical research are developed using secondary data. In its theory part, the thesis utilizes previous research about Bitcoin price formation and key investors as well as social media’s impact on decision-making and investor behaviour. Based on the findings from these researches, the thesis draws a theoretical conclusion to whether social media sentiment can influence the behaviour of Bitcoin investors, which ultimately will transfer into changes in Bitcoin price. Using Twitter as a proxy for social media, the thesis performs VADER sentiment analysis for tweets about Bitcoin collected between July 18, 2021, and July 30, 2021. The thesis then performs the Granger-causality test and vector autoregression analysis to determine whether there is a statistically significant relationship between Twitter user sentiment and Bitcoin pricing value.
Based on the thesis’s empirical findings, it seems that there is not enough empirical evidence to conclude that Twitter sentiment has a statistically significant impact on Bitcoin price. Future studies are encouraged for two main reasons. First, data with a longer timeframe or shorter frequency might yield more statistically significant results. Second, there might be another social media platform that is a better proxy for social media than Twitter.