Predicting stock returns by number of company mentions in tweets

Loading...
Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

School of Business | Bachelor's thesis

Date

2017

Major/Subject

Mcode

Degree programme

Rahoitus

Language

en

Pages

31

Series

Abstract

This study attempts to establish whether return or magnitude of return can be predicted by how many tweets mention a company by either its name or stock symbol. The sample data consists of 365 million tweets of which 706,700 mention a S&P 500 company between June 1st, 2016 and June 30th, 2017. It was found that tweets which mention a company by its stock symbol while stock markets are open have a positive impact on its return between 0 to 1%. No evidence was found of number of tweets holding a predictive value of the magnitude of return.

Description

Thesis advisor

Lof, Matthijs

Keywords

Twitter, tweet, predicting, big data

Other note

Citation