Using APIs for Data Collection on Social Media

Abstract

This article discusses how social media research may benefit from social media companies making data available to researchers through their application programming interfaces (APIs). An API is a back-end interface through which third-party developers may connect new add-ons to an existing service. The API is also an interface for researchers to collect data off a given social media service for empirical analysis. Presenting a critical methodological discussion of the opportunities and challenges associated with quantitative and qualitative social media research based on APIs, this article highlights a number of general methodological issues to be dealt with when collecting and assessing data through APIs. The article further discusses the legal and ethical implications of empirical research using APIs for data collection.

Keywords:

• APIs
• social media
• data collection
• data quality
• methodology
• research ethics

Notes

¹Even though Facebook has many public profiles that can be crawled, a majority of Facebook users have private profiles with customized privacy settings that only make content visible to a selected group of friends or limited network. In contrast, Twitter is public by default; that is, most Twitter profiles are publicly accessible online, and do not require readers to identify themselves through a login and password to access the posted tweets. As cases in this article, Facebook and Twitter thus represent a semiprivate and a public social media service, respectively. Furthermore, communication on Twitter is in a sense more distributed than on Facebook, as any given user's tweet is archived on the user's own profile, whereas on Facebook users can write on one another's profiles.

²code.google.com/p/tatiana

³www.digitalfootprints.dk

⁴Digital Footprints uses Json Array, and the data collected from Facebook were imported in the Digital Footprints local server database as multiple data points. These data were sorted according to interface categories and supplemented with different statistical functions, fuzzy search, and graphical views.