Mapping COVID-19 pandemic response apps

Project Members

Project facilitators: Anne Helmond, Fernando van der Vlist, Michael Dieter, Nate Tkacz, Jason Chao, Esther Weltevrede

Team Members: Nils Peters, Yanyi Lu, Shiman Zhang, Nil Yüce, Rita Sepúlveda, Elisavet Koutsiana, Xiaodan Feng

Table of Contents


Mobile apps are emerging as a key element in the response to the COVID-19 pandemic in countries all around the globe. This is a unique development. On the one hand, many governments and global health authorities have proposed and introduced ‘official’ apps. In some countries, their apps also immediately raised concerns related to privacy, security, adoption rates required for their effectiveness. On the other hand, there is a proliferation of third-party apps related to the pandemic. As marketplaces and distribution platforms for these apps, the leading app stores inevitably play a central role in the global response to COVID-19 at all levels of society – from government policy to the everyday practices of people worldwide.

In this project, we will map the emerging COVID-19 apps as listed in Google Play and Apple’s App Store. Although the initial focus of apps may have been on contract tracing functionality, there are now many additional apps designed in response to the pandemic, including apps for providing information, social support, symptom tracking, quarantine enforcement, and facilitating social distancing. We will scope relevant apps using a variety of relevant search queries. We look for apps in both app stores using search queries such as [COVID-19], [coronavirus], [social distancing], [quarantine], and other relevant terms and phrases associated with the pandemic.

To control this app space, app stores introduced rules and guidelines for developers using COVID-19 and related terms in app titles and descriptions. In Google Play only apps ‘published, commissioned, or authorized by an official government entity or public health organization’ may reference COVID-19 or related terms in their store listing. In addition, these apps may ‘not contain any monetizationmechanisms such as ads, in-app products, or in-app donations’ (Tolomei, 2020). The App Store similarly only allows apps from ‘recognized entities’ and prohibits ‘entertainment or game apps with COVID-19 as their theme’ (Apple Developer, 2020). This raises the question how we can locate and map the COVID-19 app space beyond the recognized entities, and which types of apps emerge in this space.

This project contributes to previous studies that began mapping COVID-19 related apps, such as reported by the Ada Lovelace Institute in their ‘COVID-19 Rapid Evidence Review’ report, the Wikipedia entry on ‘COVID-19 apps’, and a piece from the MIT Technology Review on its ‘Covid Tracing Tracker’ – a tracker for coronavirus apps. However, we adopt a broader perspective on what is and is not a COVID-19 app so that we can scope the broader variety of apps and responses. We take stock of the many COVID-19-related mobile apps available worldwide and provide a first comprehensive overview. In addition, we look at apps other than the official (governmental) COVID-19 apps, such as apps for ‘social distancing’.

Research Questions

  • How are mobile apps part of the COVID-19 pandemic response? What kind of apps have emerged – official or third-party?
  • What is the nature of the responses proposed/represented by these apps (proposed solutions, proposed technological implementations)? Which types of responses are popular? And what are the differences across countries?
  • How do app stores position themselves as part of the response? Given that Google Play is redirecting search traffic for queries such as [COVID-19], [coronavirus] to the official (governmental) apps for the user’s country, but that not all search queries are redirected, which search queries are redirected?
  • What can be said about issue frames, the actors/stakeholders involved, from these apps?


Our starting point was to query COVID-19 in the two dominant app stores, and use the app stores as demarcation devices to create our data sets (Dieter et al., 2020): Google Play and Apple’s App Store. For Google Play we immediately noticed that we were rerouted to a selected set of apps by Google in each country store. So Google is redirecting search queries to a set of approved apps per country. However, Apple takes a different approach from Google as the App Store returns ranked lists of apps for the query COVID-19. These ranked lists do not necessarily show country-specific apps on top per country.

Next, we scraped both stores for COVID-19 in all available country stores using the DMI app store scrapers using the worksheet ‘App Store Data Scraping’. For Google Play this led to a governed/redirected query collection that we analysed as the official app response per country. For Apple this led to ranked lists per country.

Actors categorisation

We took Gasser et al. (2020) as a starting point for categorising the actors behind Covid-19 apps. They distinguish between government, academic, private and citizen actors. We adjusted this by broadening ‘citizen’ to ‘civil society’, and added the category ‘public/private’ because a number of apps were developed in cooperation between public and private actors. The categories we arrived at consequently were:
  • Government
  • Academic
  • Private
  • Public/Private
  • Civil Society
In order to consistently identify the central actor behind an app, we focused on the app developer listed on the store. We clicked on each app link, read the descriptions and translated texts where necessary. If the listed developer did not clearly fall into one of the above categories, we googled for further information. Borderline and unclear cases were discussed among group members. We added an additional column (‘notes’) to include a more detailed description of the identified actor. We did not use a consistent set of terms for this. Identifying common apps in Google Play Store and App Store In order to identify apps that come from the same developer but are available in both Google Play and App Store, we first added the binary column ‘common’. 0 signifies that the app does only exist in one store, 1 signifies that the app exists in both stores. We then organised app titles alphabetically to make common apps more readily visible. Common apps were identified by answering the following three questions:
  • Do the apps have the same title?
  • Do they have the same description?
  • Do they have the same logo?
If more than one question was answered with yes, we assumed that we found a common app. To corroborate this, we checked that both apps are from different stores and whether they are from the same actor category.

Response type categorisation

We worked on all the official apps in Google Play Store under certain categories regarding their prominent features. By doing so we initially formed a dynamic taxonomy, consisting existing categories and also more specific ones emerged when the definitions were too broad or not accurately defining the apps we analyzed. We experienced this with two examples: 1) Flow modelling was a generic and neutral term for monitoring apps so we preferred to derive that into more specific categories such as quarantine compliance, contact tracing, informant, movement permit, 2) Previously came up with a category of education which we later added under information category as a part of informing the public or health care professionals for measures.

Method: While categorizing, we had gradual evaluation for each app. The description of the app was our default basis to define its category. We translated the descriptions when needed. We also worked on the visual and/or audiovisual material on the app’s page, (e.g. promotion video, photos that are showing the apps screen, translated sections when needed) for verifying the relevancy with the text or when the description of an app was insufficient for determining. When neither of these information were clear enough, in addition to checking the user comments, we also looked into tech articles or visited the developer’s webpage to collect more data on the app. The categories and their explanations are listed as following:

  • Remote healthcare: Diagnosis, prescriptions or tele-consultation to contact health services, doctor appointments
  • Symptom reporting: Comprises single interaction for reporting your symptoms through an app to an institution. This includes reporting information for test request applications.
  • Symptom tracking: Comprises usage of apps requiring daily or regularly tracking your symptom inputs.
  • Networked medicine: Comprises apps for healthcare workers to communicate and interact within a system
  • Contact tracing: Apps that are tracing who you are in contact with and notify them in case of contact with suspected cases/infected individuals
  • Quarantine compliance: Apps that “involve the real-time monitoring of whether individuals who are symptomatic or non-symptomatic are complying with quarantine restrictions.”
  • Information: comprises apps that provide official information and/or educational visual materials including symptoms, recommended safety measures, hygiene, procedures for contacting healthcare officials, quarantine, travel restrictions, government guidelines, regulations (concrete data), fact-checking (combating myths and disinformation), helpline numbers and contact details, self-assessment for symptom checklist.
  • News: comprises apps that are providing the latest news updates, notifications and up-to-date or ‘live’ information on the evolution of the virus and regional or global outbreaks of the virus, including data and statistics
  • Swiss army knife (to be decided later for the term): Apps that attempt to include “everything”, compact apps including multiple prominent feature (e.g.simultaneously containing reporting,tracking,tele-consultation,news update)
  • Hotspots: Apps giving the users real time regional data whether the location is safe from infection or not, numbers of vacant hospital beds, hospitals which provide ventilators, access to relief centres
  • Informant: Apps where users can report on the activities of others, including non-compliance with regulations, quarantine, but also excessive price increases and shortages
  • Financial aid: Apps that provide a platform for financial transactions to the people or institutions in need during the pandemic.
  • Social support: Apps that provide for requesting/contributing resources such as cooked food, shelter, instant reach for emergency lines or social networking in times of isolation or staying indoors etc.
  • Movement permit: comprising apps which have an application for movement permits, which may support a QR code scanner for security personnel to verify or QR generator for individuals to obtain for travelling within a city, region, country or crossborder.
  • Crisis communication: set up by a crisis based institution, more than mentioning coronavirus specifically, they mention disaster communication or disaster management in their description.

Technology categorisation

We do this categorization based on information of the descriptions and permissions of Covid-19 related apps in Google Play store. In total, information of 264 apps is scraped. Therefore, we split the technology research part into two sub-categories:
  • Type 1 (technology terms in translated (in English) App descriptions): GPS, Bluetooth, API, QR code, AI (artificial intellectual), Chatbot, Biometric and Cloud service.
  • Type 2 (technology terms in scraped App permission): WiFi connection information, GPS, Bluetooth and Network-based technique.
The specific technology terms are derived in the following steps: Firstly, we refer to articles (Whitelaw et al., 2020; Clearbridge Mobile, 2020), get the general ideas of what can be determined as technologies and summarize technology terms. Then we randomly read some descriptions and permissions to adjust and come to the final key terms. These keywords are filtered and labelled on Excel. In order to keep two sub-categories rigorously comparable, only these explicit and solid terms clearly mentioned in texts will be used to filter.


First, we looked into more detail into all the official apps returned per country from the Google Play store where users get redirected to the official apps for their country. We retrieved all the launch dates for all apps as well as their updates and versions.

Final Presentation – Mapping COVID-19 pandemic response apps.png

Figure A: App timeline and app responsivity (Google Play Store)

Figure A shows a timeline of the official COVID-19 apps from Google and their responsivity, their update pace. Here for the top 20 countries most affected by COVID 19 you see that a number of countries already had apps before the crisis started, these are mostly existing medical apps for communicating with health professionals that were re-appropriated for the COVID19 crisis or in the case of Brazil an e-participation app. Notably, most app development commences in early April.

Actors categorisation

Next, we examined what kind of actors are developing COVID-19 apps.


Figure 1. Actor categorisation by store

Figure 2_actors.PNG

Figure 2. Google Store Network Visualization clustered by actors categorisation

Identifying common apps in Google Play Store and App Store

The Google Play Store redirects Covid-19-related queries to ‘white-listed’, official apps.The App Store does not seem to make a distinction between official and unofficial apps. By implication, all Apple-approved apps are official apps. However, the ratio between official and unofficial apps in the App Store can be approximated by identifying app duplicates in Play Store and App Store. We found that out of 406 unique apps in the ‘Official apps (redirected search query results) – Google Play & App Store’ data set, 82 were common apps (41 in each store). Consequently, given that there are 159 apps in the App Store, 25% of these are official apps.


Figure 3. Icons from common apps between Google and Apple Stores

Response type categorisation

So what we did here to categorize the apps combining existing taxonomy and emergent categories. We looked into the descriptions of the apps, visuals and collected data through the web and user comments in case the descriptions were too short had insufficient information. We visualised our categorisation of the apps and countries. You can actually see the simplified version of categories because the apps have more than one category, so we picked the most prominent category which defines the app for the visualization. Categories are more detailed in our data set.


Figure B: Reponse types

What we found in our data set, contact tracing is quite diverse and the majority of the countries have those type of apps .When we looked into our dataset we found more Contact Tracing apps than current research by, for instance, Ada Lovelace Institute, Privacy International. And these apps are not only for personal use of individuals but also found apps for businesses an app from Singapore called SafeEntry, has a scan feature to prevent queueing to keep a record of visitors.

We also would like to briefly mention the categories for moving permit, generating QR code for roaming in a city, region or country. We found also countries that have informant apps that users can report on the activities of others, including non-compliance with regulations, quarantine, excessive price increases and shortages. And the category we called swiss army knife was quite interesting, that there are apps that provide features comprising multiple categories. Often there are many strategies overlapping within a single app.


Figure C: Response types, actors, and app permissions

Figure C includes the actors of the apps and the categorization we made and on the right are the correlation with our categories and the apps permissions. So you can see as one of the striking findings is that the informant apps, we mentioned before, informant category matches directly with the governmental apps not other actors. The other obvious point to mention with this visualisation is that the majority of apps are being produced by the government, but there are private actors involved. Private actors are the primary group developing bluetooth contact tracing.

Technology categorisation

Finally, we examined the distribution of used technologies and searched the app descriptions for mentions of technologies. As can be seen from Figure 4, GPS is the most frequently mentioned technology, followed by Bluetooth and network.


Figure 4. Comparison of information contained between sub-categories

We find that 83.40% of the apps do not explicitly mention technology used (Figure 5), while in permission information, only 20.24% among them do not explicitly mention the technologies being used (Figure 6).


Figure 5. Apps explicitly mention technological terms in description


Figure 6. Apps explicitly mention technological terms in permission

Figure 7. Comparison of explicit mentions between sub-categories.jpg

Figure 7 shows the number of apps which the description and the permission both mention, neither mention, only description mentions and only permissions mentions. The highest purple column with a number over 160 corroborates the above finding.

Interpretation and Case study

The finding that many apps do not explicitly mention technology in their descriptions can be interpreted as follows. In Google Play store, the descriptions of apps are not regulated. In other words, the platform does not enforce app developers to include specific information. As a result, the developers are freely to choose what kind of information they would like to provide or even to provide nothing. On the other hand, the permissions are much more regulated. In the details, we can clearly find what kind of settings of our phones will the app has access to, location, camera, storage, ect. The app descriptions are often framed in a way of service they are able to provide, in other words, strong points stimulating people to download other than technologies that maybe some people concern as an invasion of privacy like discussed in How do COVID-19 tracing apps work and what kind of data do they use?, while the permission information is more factual and complete in disclosure.

This large difference of technological information contained between the app descriptions and permission information can say more. Critical discourse/textual analysis could be done in those app descriptions for the attempt to further elaborate on the above point. For example, an app called “Coronika - Your corona diary” frames its app descriptions largely with information tailoring to the audience: “helps you remember who you met and where you have been- to reduce the spread of the virus” It says the app is “good for you” because: It stresses the “data remain locally” multiple times, while very much understates and slides over the fact that this is a private-developed app serving the need of the government by saying “It is essential for health authorities to understand where infected people have been in order to identify sources of infection and contact people.”

Discussion and conclusion

We have found different types of governance and organisational logic in relation to COVID-19 by the two prominent app stores: from query rerouting by Google to offering ranked lists by Apple.

We have created a complete data set of ‘official responses’ per country that contributes to existing lists such as the Wikipedia entry on ‘COVID-19 apps’, and a piece from the MIT Technology Review on its ‘Covid Tracing Tracker’ – a tracker for coronavirus apps. In this set we found that beyond governmental apps there are many private sector (start-up) collaborations. For example, in Brazil there is an app by a start-up that seems to be ‘stepping in’ in lack of a governmental response.

Overall we have found that the apps offer different types of approaches: from the Swiss-army-knife solution of fitting all kinds of different solutions (remote healthcare, symptom reporting, contact tracing, movement permit, etc) into a single app, to single-purpose apps that facilitate quarantine enforcement. To some extent we did not find the discourse of ‘technological solutionism’ in the app descriptions as only few apps explicitly mention APIs, protocols, or other technologies used.



Topic attachments
I Attachment Action Size Date Who Comment
Figure 2_actors.PNGPNG Figure 2_actors.PNG manage 351 K 15 Jul 2020 - 11:19 AnneHelmond  
Figure 7. Comparison of explicit mentions between sub-categories.jpgjpg Figure 7. Comparison of explicit mentions between sub-categories.jpg manage 34 K 15 Jul 2020 - 11:26 AnneHelmond  
Figure1_actors.PNGPNG Figure1_actors.PNG manage 77 K 15 Jul 2020 - 11:19 AnneHelmond  
Figure3_actors.PNGPNG Figure3_actors.PNG manage 247 K 15 Jul 2020 - 11:24 AnneHelmond  
Figure4_technology.pngpng Figure4_technology.png manage 22 K 15 Jul 2020 - 11:25 AnneHelmond  
Figure5_technology.pngpng Figure5_technology.png manage 38 K 15 Jul 2020 - 11:25 AnneHelmond  
Figure6_technology.pngpng Figure6_technology.png manage 45 K 15 Jul 2020 - 11:25 AnneHelmond  
FigureB-responsetypes.pngpng FigureB-responsetypes.png manage 242 K 15 Jul 2020 - 12:09 AnneHelmond  
FigureC-responsetypes2.pngpng FigureC-responsetypes2.png manage 1 MB 15 Jul 2020 - 12:10 AnneHelmond  
Final Presentation – Mapping COVID-19 pandemic response apps.pngpng Final Presentation – Mapping COVID-19 pandemic response apps.png manage 143 K 15 Jul 2020 - 11:56 AnneHelmond  
Topic revision: r1 - 15 Jul 2020, AnneHelmond
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback