Health SDI, Simple APIs and Web Map Clients for COVID Tracking

Sharing of health data has become one of the most important data sharing activities in the last months. The general population and authorities need to know the current situation not only about reported cases, deaths and recovered, but other data (e.g. hospital beds, ventilators, etc.) that supports properly responding to the COVID-19 pandemic. In addition to this health data, there are other non-health datasets (e.g. boundaries, socio-economic data) that provide context, complement and help decision makers better understand the possible actions to take. So, in today’s health emergency situation, an information sharing infrastructure with modern web visualization tools is critical to guide the way we share health and complementary data. 

In this blog I am going to 1) highlight the importance of a Health Spatial Data Infrastructure (SDI) documented in a recent paper published by OGC, which I co-authored, 2) make the case that sharing of data should be easy, and 3) provide an example of a dashboard tracking COVID-19 data based on MapStore.

The Open Geospatial Consortium (OGC) published some weeks ago the white paper: Health Spatial Data Infrastructure: Application Areas, Recommendations, and Architecture. It provides reference to initiatives (e.g. EO4Health, Eo2Heaven, INSPIRE Human Health and Safety Specifications,)  that have occurred in the last years related to building an infrastructure that enables the collection, exchange, integration, analysis, and visualization of health and non-health data to help identify and address health issues at global and local level. The sections I consider worth reading are: Section 4.7. Pandemic Response,  Section 5 Data Considerations and Related Recommendations, and  Section 6 – Health SDI Architecture Framework.

Section 4.7 discusses the development and use of different indexes (e.g. Transmission Risk Index) that are only possible to develop if there is enough information available. This is the reason why a Spatial Data Infrastructure is critical. This is the reason why we need to publish the data on the Web using open standards, following recommendations from Section 5 that takes into account security, anonymization and suggested formats and interfaces (e.g. WFS or WMS). 

Section 6 provides a discussion on the architecture and also a workflow about what is needed when creating indexes that involve different types of data (See Figure 1). What is critical from the workflow is the catalog, which is the heart of an SDI. If you don’t know what data is available you can’t put it together in a fast manner and it requires phone calls, emails, searching, scraping websites, etc., wasting precious time when building a common operational picture.

Figure 1 – Health SDI Workflow

Worth to mention, that the prototype of a Health SDI was advanced in a recent OGC initiative, well documented in an OGC video. The tool used was GeoNode which is an-easy-to-setup tool for SDIs that allows the data to be easily cataloged, updated and shared via open standards.

Organizations harvesting information from different sources to create dashboards mostly rely on personal communications and getting the data from official web reports available at government/intergovernmental websites. Then, they create “machine readable formats” such as JSON or CSV that are ingested in the web clients. The “manual process” of getting the data requires a lot of human power, and fortunately for this crisis there are a lot of people willing to help. This is not the ideal. Government and other official sources should be making data available via open standards following the recommendation in the report.

At GeoSolutions, we were investigating how COVID-19 data was being published. We found an initiative, the COVID Tracking Project that actually does a great job on aggregating data for the US states. They perform quality control, publish the provenance, and make the data available via a simple Data API. If organizations around the world can follow simple APIs like the one mentioned, it will be easier to aggregate data in our current and other health emergency events. Also, fortunately, OGC is working on simple APIs that will enable government, agencies and projects (similar to the COVID Tracking Project), to share data via an open-standard simple API

Based on the previous mentioned project, and the simplicity of their API, we wanted to develop a simple sleek dashboard for tracking COVID-19 cases in the US. We wanted to reuse the best tools that we have in house at GeoSolutions, so we reused MapStore to build such a dashboard; you can reach it live here.

Figure 2 – US COVID Tracking Dashboard with MapStore

The client interacts directly with the API. There is no database on the backend and this is the reason it can be installed in a S3 bucket (which is a cheaper option than using a virtual machine under EC2). The user interface allows to view more than one variable at a time by enabling the user to select from the left side and displaying bubble plots located at the centroids of each state. The user can also view and order the statistics on the right column enabling easy comparison among the states. Hovering over a state shows all the data for that particular state, as it is shown for New York . Also, in our spirit of open source advocates, we have released the code at GitHub under a BSD License.

Sharing data via open standards is important, as well as, simple APIs that can enable web clients to provide dashboards and critical information to end users. I hope you find this blog useful and can inspire some ideas about sharing and visualizing pandemic tracking.  If you would like GeoSolutions to help with your health data publication, cataloging or visualization needs we are here to help!

Please don’t hesitate to contact us. We would love to hear how we can help you!

Cordially,

Luis Bermudez

CEO GeoSolutions USA