Title: Data modelling for large-scale social media analytics: design challenges and lessons learned

Authors: Ahmet Arif Aydin; Kenneth M. Anderson

Addresses: Department of Computer Science, Inonu University, 44280, Malatya, Turkey ' Department of Computer Science, University of Colorado, Boulder, CO 80309-0430, USA

Abstract: We live in a world of big data; organisations collect, store, and analyse large volumes of data for various purposes. The five V's of big data introduce new challenges for developers to handle when performing data processing and analysis. Indeed, data modelling is one of the most challenging and critical aspects of big data because it determines how data will be structured and stored; these decisions then impact how that data can be processed and analysed. In this paper, we report on designing a data model for storing and analysing Twitter data in support of crisis informatics. In this work, we leverage the data model provided by columnar NoSQL data stores to design column families that can efficiently index, sort, store and analyse large Twitter datasets. In particular, our column families are designed to achieve efficient batch data processing. We evaluate these claims and discuss our future work.

Keywords: data modelling; social media analytics; big data analytics; NoSQL.

DOI: 10.1504/IJDMMM.2020.111409

International Journal of Data Mining, Modelling and Management, 2020 Vol.12 No.4, pp.386 - 414

Received: 11 Dec 2019
Accepted: 13 Jun 2020
Published online: 25 Nov 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article