A PLUS ON THE DEEPMAP COVID-19 DASHBOARD
There is a great deal of disruption in the Netherlands. The daily lives of millions of Dutch citizens have drastically changed in a short period of time. The corona crisis is exerting tremendous pressure on our well-organised healthcare system and our economy. However, it is precisely in times like this that social initiatives designed to help the country and its fellow citizens often emerge. This phenomenon is also evident within the data science community, which is hard at work with the open source data of various government and other organisations, such as the Netherlands National Institute of Public Health and Environmental Protection (RIVM), Statistics Netherlands (CBS) and the National Intensive Care Evaluation (NICE) Foundation. Government and other organisations are making their data available to enable citizens and organisations to analyse it and use it to conduct research. This can result in many benefits to government and society. The benefits vary from increasing the transparency of government actions and accountability, stimulating economic and social innovation, and contributing to a more effective and efficient government.
During these extraordinary times, the relevance of data in developing an effective health policy that prevents further spread is evident. Ynformed naturally would like to make a contribution and this is why it has organised a digital COVID-19 hackathon in cooperation with Sjoerd Wierenga, founder of DEEPMAP. This blog describes the results.
DEEPMAP Municipal Health Service (GGD) COVID-19 Dashboard
Sjoerd Wierenga recently approached Ynformed to contribute to its initiative: the GGD COVID-19 Dashboard. This interactive dashboard helps Municipal Health Service (GGD) regions to gain insight into the current status of COVID-19 developments within all GGD regions. The dashboard consolidates various open source data sources provided by the RIVM, CBS and the NICE Foundation, and transforms this data into valuable information. This is an example of how the data science community can contribute in providing professionals and policymakers within the GGD with current insights and trends. Naturally with a view to promoting data-driven policy formulation.
Getting down to work – the COVID-19 Hackathon
With a small group of colleagues, we started working on the question raised by Ester de Jonge of the Province of South Holland South GGD region. She informed DEEPMAP that she would like to know whether the speed with which the established number of infections is spreading differs between densely and less densely populated areas. But how do you establish this speed when it is constantly changing? Different approaches were explored that of course also produced different results.
First, we analysed the available COVID-19 data. This made it clear that concepts such as ‘doubling time’ and ‘growth factor’ are complex because these concepts are directly influenced by changes in policy, for example. Due to these policy changes, the growth factor of confirmed infections can suddenly significantly increase or decrease. This directly affects the subsequent analyses.
Next, we observed a number of interesting correlations. For example, we found a correlation between the speed of the spread (expressed as the average doubling time) and the population density at the municipal level (expressed as the Surrounding Address Density (SAD)). In addition, there appear to be correlations between spread and health indicators. These indicators are described in the health monitor carried out in 2016, and among other things concern smoking behaviour.
We are currently working on further elaborating on these insights. Following a validation round, the analyses may be given a spot on the dashboard, after which they can be directly consulted in the open source dashboard. An end condition is that the analyses must provide clarity without generating too many new questions (which in turn would require explanation). This will be confirmed after the validation round with domain experts.
Interested in contributing to the dashboard? For example by giving substantive feedback, performing data analyses, reviewing code or something else? If so, contact Sjoerd Wierenga or Theodor Poels.