There have been recent efforts to use social media to estimate demographic characteristics, such as age, gender or income, but there has been little work on investigating the effect of data acquisition methods on producing these estimates. In this paper, we compare four different Twitter data acquisition methods and explore their effects on the prediction of one particular demographic characteristic: occupation (or profession). We present a comparative analysis of the four data acquisition methods in the context of estimating occupation statistics for Australia. Our results show that the social network-based data collection method seems to perform the best. However, we note that each different data collection approach has its own benefits and limitations.
Paper
Full text
The Effects of Data Collection Methods in Twitter
Semantic Scholar · Computer Science · 2016
Abstract
There have been recent efforts to use social media to estimate demographic characteristics, such as age, gender or income, but there has been little work on investigating the effect of data acquisition methods on producing these estimates. In this paper, we compare four different Twitter data acquisition methods and explore their effects on the prediction of one particular demographic characteristic: occupation (or profession). We present a comparative analysis of the four data acquisition methods in the context of estimating occupation statistics for Australia. Our results show that the social network-based data collection method seems to perform the best. However, we note that each different data collection approach has its own benefits and limitations.