RIP Twitter API: A eulogy to its vast research contributions

Autor: Murtfeldt, Ryan, Alterman, Naomi, Kahveci, Ihsan, West, Jevin D.
Rok vydání: 2024
Předmět:
Druh dokumentu: Working Paper
Popis: Since 2006, Twitter's Application Programming Interface (API) has been a treasure trove of high-quality data for researchers studying everything from the spread of misinformation, to social psychology and emergency management. However, in the spring of 2023, Twitter (now called X) began changing $42,000/month for its Enterprise access level, an essential death knell for researcher use. Lacking sufficient funds to pay this monthly fee, academics are now scrambling to continue their research without this important data source. This study collects and tabulates the number of studies, number of citations, dates, major disciplines, and major topic areas of studies that used Twitter data between 2006 and 2023. While we cannot know for certain what will be lost now that Twitter data is cost prohibitive, we can illustrate its research value during the time it was available. A search of 8 databases and 3 related APIs found that since 2006, a total of 27,453 studies have been published in 7,432 publication venues, with 1,303,142 citations, across 14 disciplines. Major disciplines include: computational social science, engineering, data science, social media studies, public health, and medicine. Major topics include: information dissemination, assessing the credibility of tweets, strategies for conducting data research, detecting and analyzing major events, and studying human behavior. Twitter data studies have increased every year since 2006, but following Twitter's decision to begin charging for data in the spring of 2023, the number of studies published in 2023 decreased by 13% compared to 2022. We assume that much of the data used for studies published in 2023 were collected prior to Twitter's shutdown, and thus the number of new studies are likely to decline further in subsequent years.
Comment: 16 pages, 4 figures, 4 appendices
Databáze: arXiv