Text mining applications to support health library practice: A case study on marijuana legalization Twitter analytics

Abstract

Background

Twitter is rich in data for text and data analytics research, with the ability to capture trends.

Objectives

This study examines Canadian tweets on marijuana legalization and terminology used. Presented as a case study, Twitter analytics will demonstrate the varied applications of how this kind of research method may be used to inform library practice.

Methods

Twitter API was used to extract a subset of tweets using seven relevant hashtags. Using open-source programming tools, the sampled tweets were analysed between September to November 2018, identifying themes, frequently used terms, sentiment, and co-occurring hashtags.

Results

More than 1,176,000 tweets were collected. The most popular hashtag co-occurrence, two hashtags appearing together, was #cannabis and #CdnPoli. There was a high variance in the sentiment analysis of all collected tweets but most scores had neutral sentiment.

Discussion

The case study presents text-mining applications relevant to help make informed decisions in library practice through service analysis, quality analysis, and collection analysis.

Conclusions

Findings from sentiment analysis may determine usage patterns from users. There are several ways in which libraries may use text mining to make evidence-informed decisions such as examining all possible terminologies used by the public to help inform comprehensive evidence synthesis projects and build taxonomies for digital libraries and repositories.