Learn to Rehydrate Twitter Data Using Python: A #hellobrother Case Study (2019)

Abstract

In research, it is common practice to share Twitter datasets using only tweet identifiers; this is done by way of dehydration and the subsequent rehydration of the dataset for further qualitative analysis. This tutorial demonstrates the process of rehydration that simply refers to using the Twitter Application Programming Interface to recollect or retrieve the tweets that the tweet identifiers refer to using Python programming language, Jupyter Notebooks, and a third-party tool named Twarc. Rehydration is the standard way to share Twitter data in accordance with Twitter Terms of Service that only allows the sharing of tweet identifiers, not the full tweet data. In addition, this tutorial explores the degradation and disappearance of data that occurs when tweets are removed by users or through moderation and which becomes evident once the data is rehydrated and not all of the dataset can be retrieved. Rehydration is especially useful for mixed-method approaches that include qualitative ethnographic analysis of computational data because rehydration not only allows the sharing of datasets between researchers but also enables the (re)construction of the field for ethnographic analysis. The tweetset under investigation in this tutorial includes tweets tagged with the commemorative hashtag #hellobrother, collected in the context of the Christchurch mosque attacks in 2019. Although the #hellobrother tweetset is mainly commemorative in nature, some content might be perceived as more sensitive. This tutorial includes the #hellobrother tweetset in the form of tweet IDs, a How-to Guide for rehydrating the data, as well as an analytical notebook for exploring the rehydrated tweetset.