Issues of Scale in the Corona Crisis

The granularity at which you look at COVID-19 may determine your attitude towards Sars-CoV-2

Scale is one of the most fundamental concepts in Geography. My PhD student just completed her dissertation on “The Consequence of Scale: Process and Policy Implications of Composite Index Modelling Using the Conceptual Framework of GIS-MCDA”, in which she compares biodiversity indices computed at different scales within a city, for example smaller census tracts vs larger social planning neighbourhoods. In Geographic Information Systems (GIS), we usually work with aggregated data, and the scale of aggregation can range from census blocks through postcode areas and neighbourhoods/wards to cities, counties, provinces, and countries. Results of data analytics are known to depend on several aspects of scale, including the observation/measurement scale, at which data are collected; modelling scale, at which data are analyzed; and operational/policy scale(s), at which decisions are made and implemented.

Geographic scales play an important role in the ongoing corona crisis too. Sars-CoV-2 and COVID-19 statistics are being mapped by country at the global scale and by state or province at the national scale. You will also find within-city maps such as the “City of Toronto COVID-19 Summary” by neighbourhood shown below, which exemplifies two typical communication errors:

  • Choropleth maps must never be used to map raw-count data such as COVID-19 cases, since raw-count data depend on the sizes of the underlying spatial units in terms of geographic area or total population, while the cartographic symbol (colour shading) conceals this dependency, see my March 26th blog post at
  • Toronto Public Health’s map conflates cases in the community with those from institutional outbreaks, which are assigned to the neighbourhood that the institution (e.g. longterm-care home) is located in, while the processes leading to transmissions and the resulting policy decisions are not comparable, see Chris North’s June 16th analysis of an earlier version of this map at
City of Toronto COVID-19 summary application with map of cumulative cases by neighbourhood. Note that this is a misleading use of a choropleth map for raw count data. In addition, this view conflates localized institutional outbreaks with community spread of Sars-CoV-2. Source:

In addition to refining our scales of observation, analysis, and policy-making when it comes to institutional settings such as longterm-care homes or schools, much of the medical research around COVID-19 is scaled further down to the individual person’s level or to even more granular detail such as the cell level. While trying to find commonalities within groups of subjects, the researchers examine individual responses of the human body to the virus; mechanisms of individual transmission or immunity; or people’s behavioural responses to emergency orders, to name a few examples. In contrast, epidemiologists will generally study populations using aggregated data and/or mathematical models to describe issues like disease spread and predict its future development.

With COVID-19, as with many other news topics, there are numerous threatening findings and personal horror stories to choose from. It is my impression that the media reporting and images focus on local hardship, such as an individual with extended illness or an overloaded hospital or an outbreak in a neighbourhood school. The more benign overview of the state of the pandemic is not presented with the same intensity, and in addition, metrics, graphs, and maps are often based on raw counts and cumulative data – in essence the combination of the individual stories – rather than putting the counts into perspective. I have already written about the lack of data normalization and issues with classifying data such as the deaths from, or with, COVID-19. Maps in particular may not support a proportionate situation assessment, as their appearance depends largely on value ranges and the cartographer’s classification and symbol choices.

Individual stories are typically negative, since newspaper readers and TV viewers do not seem interested in positive news. Recipients who prefer concrete stories over abstract data will therefore perceive only negative information. This bias likely includes journalists who then amplify this perspective by focusing on cautious, risk-averse expert opinions confirming simplistic messages such as “every life counts”. I therefore contend that the scale of perception, at which we consume information about the crisis, largely determines our attitude towards the virus and towards societal response measures such as lockdowns, distancing, and mask-wearing. In other words, people who perceive the pandemic through the dramatic videos from Northern Italy or New York City hospital chaos, focus on atypical death reports such as those of children dying, or read stories about #LongCovid, will call for new lockdowns, defend their physical distance, or report partying neighbours out of fear for their own or others’ health. Meanwhile, people who look at the pandemic at a coarser scale of epidemic curves or normalized test positivity rates or death rates will more likely realize that the situation is not quite as dire as initially predicted and adjust their personal response accordingly.

Coincidentally, the Ontario Minister of Long-Term Care, Dr. Merrilee Fullerton, was asked in parliament today about a Toronto Sun column comparing deaths from COVID-19 to Influenza, and is quoted in QP Briefing:

“The numbers do indicate — if you actually measured the flu season, from 2017 into 2018 — the numbers are comparable. But I don’t want to talk about numbers. You know, it is about people.”

Now, while society and politics sure are about people, important decisions are usually made based at least in part on data! Thus, the Minister’s statement would suggest that we do not act much differently from managing the annual flu cycle. Nevertheless, all bets are off on an imminent announcement of new restrictions on indoor dining and certain group activities due to perceived “spike” in COVID-19 “cases” in Ontario. Nevertheless, I am still hopeful that a larger scale of perception will prevail and the government stick to a moderate approach balancing health and prosperity, as argued for example in May in an open letter from a mixed group of experts with the MacDonald-Laurier Institute.