Application of Matrix Multiplication Technique to Google Trends Data for Mining Knowledge on Nigeria’s Telecommunications Industry

Abstract

Adeola O. Opesade

Africa Regional Centre for Information Science, University of Ibadan, Nigeria

ao.opesade@ui.edu.ng, morecrown@gmail.com

Abstract

Data is the raw material of the present Information Age. While there are many sources of big data, the rapid growth of the web and the variety of its data types has made it the largest publicly accessible data source in the world. Google Trends (GT), a web-based data source, has been investigated and analysed by many previous studies. It could however, be observed that these previous studies have mostly analysed GT data based on either time or geographical locations. The present study applies the mathematical principle of matrix multiplication to extend the use of GT data for mining purposes. Dataset derived from the proposed data integration model was evaluated on Nigeria's Global System of Mobile Communication (GSM) operators' (MTN, AIRTEL, GLO, ETISALAT) timeline and geographical search volumes. Supervised learning experiment on the integrated dataset resulting from the proposed model performed very favourably. Further analyses on the resulting dataset showed that search volume for MTN was the most consistent across the 37 Nigerian locations. MTN search also had the highest coverage while AIRTEL had the highest average search volume among the four operators. Analysis of Variance and Post-Hoc tests showed that the search volumes of the four GSM operators were significantly different from one another.