The Predictive Power of Social Media Data

Publikation: Bog/antologi/afhandling/rapportPh.d.-afhandling

583 Downloads (Pure)

Abstract

Predictive analytics using social media data is a relatively new field, being widely recognized at first due to “Predicting the Future with Social Media” of Asur and Huberman (2010). Their article proposed a model that could predict the revenues of Hollywood movies using Twitter data as input. This model has proven more accurate than the golden standard for movie revenue prediction – the Hollywood Stock Exchange – being thus a betting market for Hollywood movie performances and considered a good prediction market example.
Since 2010, predictive analytics using social media data have developed extensively. In line with this trend, this PhD thesis presents pioneering research on the first Twitter-based prediction sales model, which explains why Twitter data can predict iPhone sales. It also demonstrates that similar predictive sales models can be built using Facebook and Google search data. The methodology is based on customer journey models, which are the main conceptual models for categorizing all social media and web search data in a sales context. Based on this methodology, an investor journey model for categorizing web search data in a financial market context is proposed.
OriginalsprogEngelsk
UdgivelsesstedFrederiksberg
ForlagCopenhagen Business School [Phd]
Antal sider327
ISBN (Trykt)9788775681990
ISBN (Elektronisk)9788775682003
DOI
StatusUdgivet - 2023
NavnPhD Series
Nummer29.2023
ISSN0906-6934

Citationsformater