Lead: Anne Helmond, Fernando van der Vlist, Esther Weltevrede (alphabetical)
Participants: Carolin Gerlitz, Shefali Bharati, Taylor Geiger, Ine van Zeeland, Maggie MacDonald, Stephanie de Smale, Emanuela Blaiotta, Maria Fernanda Ibañez Duarte, Julia Wolny, Janna Joceli Omena, Ana Pop Stefanija, Christina Meyenburg, Michael Dieter, Jason Chao, Nate Tkacz, Serena Del Nero, Marco Mezzadra
Mobile apps have become a popular new cultural medium. Today, the main entry point to these apps – for developers and for users alike – is via one of the app stores, where users can search for the name of individual apps (e.g. [Moodpath]) or put query to demarcate collections and genres of apps (e.g. [depression]). However, little is known about the ways in which these search results are constructed or how related app recommendations emerge. What is app relatedness within the specific setting of app stores? At first glance, related apps seem to have a topical coherence with the source app. Upon closer inspection we can ask: why or when is a mindfulness app, for example, related to a depression app? And how do larger mindfulness app spheres relate to depression app spheres? What agencies produce relations among apps (e.g., ‘You might also like’, ‘Similar apps’), and how might these algorithmic devices be repurposed? What kinds of recommendations do app stores suggest for controversial apps or apps with sensitive content (e.g., abortion apps, religion apps)? For example, when do app stores recommend bible verses or a ‘starter-kit with pro-life knowledge’ when searching for [abortion] – do app stores have a ‘pro-life bias’ (Devaney, 2016)? So more broadly, how might researchers reverse-engineer app store results and recommendations, and to what extent might researchers repurpose them for social and/or medium research?
- How to reverse-engineer app store results for social and/or medium research?
- How to reverse-engineer (the bias of) app store recommender algorithms?
- How to compare the classification and recommendations of apps in Google Play (Alphabet) and iTunes Store (Apple)?
- How is bias in app stores addressed socially? How do the media, journalists, developers, and others discuss the bias of app stores differently?
Part I: ‘There’s an app for that’
Part II: There’s no app for that
Part III: Crappy stores
- Balashankar A, Koc L, & Guimaraes N (2016, December 14) App Discovery with Google Play, Part 2: Personalized Recommendations with Related Apps. In: Google AI Blog. Available from: http://ai.googleblog.com/2016/12/app-discovery-with-google-play-part-2.html.
- Devaney B (2016) Does Apple have a problem with abortion? The App Store pro-life bias. Available from: https://www.gadgette.com/2016/01/05/does-apple-have-a-problem-with-abortion-the-app-store-pro-life-bias/.
- Haldar M, MacMahon M, Jha N, et al. (2016, November 8) App Discovery with Google Play, Part 1: Understanding Topics. In: Google AI Blog. Available from: https://research.googleblog.com/2016/11/app-discovery-with-google-play-part-1.html.
- Lee H-C, Chen X, and An Q (2017, January 30) App Discovery with Google Play, Part 3: Machine Learning to Fight Spam and Abuse at Scale. In: Google AI Blog. Available from: https://ai.googleblog.com/2017/01/app-discovery-with-google-play-part-3.html.
- Rahman M (2018, June 27) Developers are facing huge drop in new installs after Play Store algorithm changes. In: XDA Developers. Available from: https://www.xda-developers.com/developers-huge-drop-new-installs-play-store-algorithm-changes/.
- Rieder B, Matamoros-Fernández A and Coromina O (2018) From ranking algorithms to ‘ranking cultures’: Investigating the modulation of visibility in YouTube search results. Convergence 24(1): 50–68. DOI: 10.1177/1354856517736982.
- Weltevrede E (2016) Google algorithm changes and the volatility of method. In: Repurposing digital methods: The research affordances of platforms and engines. Ph.D. thesis, University of Amsterdam. Available from: http://hdl.handle.net/11245/1.505660.