Coronavirus and population

<a href="https://storymaps.arcgis.com/stories/be6103d0004e4411929e668d20a55883" rel="noopener noreferrer" target="_blank">Sea-level rise</a> studies for local climate initiatives and <a href="https://storymaps.arcgis.com/stories/a5621cbb966c41edba927024b4ee34ea" rel="noopener noreferrer" target="_blank">new map techniques</a> on older projects also put me in touch with Esri's John Nelson and Kenneth Field.

The Coronavirus outbreak brought many resources from <a href="https://www.esri.com/en-us/covid-19/overview" rel="noopener noreferrer" target="_blank">Esri, Inc.</a> as well as <a href="https://covid19.esriuk.com/" rel="noopener noreferrer" target="_blank">Esri(UK)</a>. I created this dashboard using <a href="https://www.gov.uk/government/publications/covid-19-track-coronavirus-cases" rel="noopener noreferrer" target="_blank">Public Health England</a> (PHE) and <a href="https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/articles/coronaviruscovid19roundup/2020-03-26" rel="noopener noreferrer" target="_blank">Office for National Statistics</a> (ONS) data.

Second is the identification of youngest (white) and oldest (yellow) age tiers, which are the most vulnerable to COVID-19.

All linked in text unless otherwise stated:

Inspired by PHE <a href="https://www.arcgis.com/apps/opsdashboard/index.html#/f94c3c90da5b4e9f9a0b19484dd4bb14" rel="noopener noreferrer" target="_blank">desktop</a> and <a href="https://www.arcgis.com/apps/opsdashboard/index.html#/ae5dda8f86814ae99dde905d2a9070ae" rel="noopener noreferrer" target="_blank">mobile</a> dashboards and many others on <a href="https://storymaps.arcgis.com/stories/4fdc0d03d3a34aa485de1fb0d2650ee0" rel="noopener noreferrer" target="_blank">ArcGIS Online</a>, I adapted data sets available in England for East Anglia. While map data are dynamic, this dashboard is updated manually every day.

Enter the latter's <a href="https://www.esri.com/arcgis-blog/products/arcgis-pro/mapping/mapping-coronavirus-coxcombs/" rel="noopener noreferrer" target="_blank">Mapping coronavirus coxcombs</a> with attendant caveats, such as clearly explaining it as on <a href="https://carto.maps.arcgis.com/apps/webappviewer/index.html?id=7f030ac941e74891a394cfaa5ea852cb" rel="noopener noreferrer" target="_blank">his map</a>:

You will however notice these symbols overlap: So how to compare cases with age groups, other than toggling them on&amp;off on the web map?

... But wait! Might there be other data sets to possibly post like so?

Another caveat was to use a proper scale &amp; projection, so these symbols plot correctly - British National Grid was used here, by first posting a County vector tile in said projection - this is discussed in detail under A word on tech at the end of <a href="https://storymaps.arcgis.com/stories/a5621cbb966c41edba927024b4ee34ea" rel="noopener noreferrer" target="_blank">this story map</a>.

Note also that only case numbers are tallied historically by PHE - new and cumulative cases and deaths (recoveries not after 22 Mar 2020) are posted only daily, to be collated someday unless PHE does it first - so we don't have the richness WHO have by country.

Enter the Story map swipe and spyglass web app shown here - the condensed version of this story map if you will - enjoy!

"So much information in one go. Each time I look there's some new information to consider..."?

As Kenneth Field said "The first task (as with most mapping projects) is getting your data into shape" - I spare you the wrangling to get here - so I aggregated ONS table into age classes, classed them as above, joined them to Unitary Authority centroids, ran the coxcomb GP tool... et voila!

First I adapted Public Health England coronavirus case numbers per Unitary Authority (the smallest area UK government releases such data) for 3 - 29 March 2020. Note that greater London area was omitted for clarity, it can be restored by removing the filter, if you clone:

<li>this attempts to use clever symbology and display many data dimensions in one go<li>that in turn allows the comparison of separate data sets by simple overlay &amp; swiping</li><li>all to help see relationships between disparate data sets and draw further conclusions</li></li>

So after many online discussions on the pitfalls of using proper <a href="https://medium.com/zolnai-ca/carbon-footprint-through-time-fec2e3b7c6ca" rel="noopener noreferrer" target="_blank">styling</a> and <a href="https://medium.com/zolnai-ca/carbon-footprint-through-time-ii-d62ce3a1b6ae" rel="noopener noreferrer" target="_blank">projection</a>:

Until self-isolation I helped with local <a href="https://www.cottenhampc.org.uk/cottenham-covid-19-support-team/" rel="noopener noreferrer" target="_blank">Parish Council</a> and Extinction Rebellion Cambridge <a href="https://www.scientistsforxr.earth/" rel="noopener noreferrer" target="_blank">Scientists</a>. So I started blogging a <a href="https://blog.zolnai.ca/2020/03/coronavirus-update.html" rel="noopener noreferrer" target="_blank">Coronavirus Update</a> page. It was natural then to post more details on this dashboard as <a href="https://blog.zolnai.ca/2020/03/local-coronavirus-support.html" rel="noopener noreferrer" target="_blank">Local Coronavirus support - updated daily</a>.

First to jump out, is what I call "reverse coxcombs": they spiral counterclockwise compared with clockwise above. It helps contrast them by shape, whilst keeping the colour scheme the same.

This is a WhatsApp message posted earlier to a local scientists' group in Cambridge UK. It belies both the challenges and opportunities of the current lock-down. But how did we get to the banner map, which as <a href="https://twitter.com/John_M_Nelson" rel="noopener noreferrer" target="_blank">@john_m_nelson</a> tweeted about the original map that inspired me in the first place:

Enter ONS population estimates by age for mid-2018 on 2019 Local Authority boundaries. They <a href="https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/articles/wheredoover70slive/2020-03-24" rel="noopener noreferrer" target="_blank">post here</a> their beautiful but unlinked map of over-70s by post code, but the linked data sets are incomplete, so a little sleuthing uncovered <a href="https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernireland" rel="noopener noreferrer" target="_blank">a usable set</a>.

While Unitary Authority area centroids are easy to create, what about their populations by age? Coxcombs work on timelines, so eventually I hit upon spoofing ages as dates and restoring the attributes in the legend like so:

Comparing pandemic cases and population age groups in reverse coxcomb symbols (cases increase with time, pop. groups decrease with age) highlights vulnerable younger and older