Abstract

Based on an analysis of several sampled political discussions on Twitter, a new methodology is proposed that allows researchers to obtain a significant, replicable, and manageable data sample from a universe of Twitter metadata. The proposal is a new model called Top discussion indicator (TDI). The aim of TDI is to assist researchers in obtaining a representative set of text from Twitter that includes the minimum amount of information needed to generalize the results.

Abstract

Based on an analysis of several sampled political discussions on Twitter, a new methodology is proposed that allows researchers to obtain a significant, replicable, and manageable data sample from a universe of Twitter metadata. The proposal is a new model called Top discussion indicator (TDI). The aim of TDI is to assist researchers in obtaining a representative set of text from Twitter that includes the minimum amount of information needed to generalize the results.