You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
1/ Grab a cup of coffee (or tea)
A comprehensive, neighborhood-by-neighborhood review of #SARSCov2 prevalence/trends in the City of Toronto.
% positivity & cases, with weekly trends since Aug, AND:
*cross referenced with neighborhood census data*
The findings are incredible.pic.twitter.com/MPxafNDLi2
2/ Note: even if you are not in Toronto/Canada, I think you will find this data/analysis compelling, and universally applicable re #SARSCov2/#COVID19 learning and public policy implications.
Toronto’s diversity (>51% visible minority) makes it an interesting case study.pic.twitter.com/aZX2iiqeUe
3a/ In this thread, I show/illustrate: 1. for Toronto’s 140 neighbourhoods (and groups of hoods, e.g. DT Core, Northwest), which have increasing/decreasing % pos & cases per 100k. (Some peaked long before the Oct 10th restrictions. Others still increasing despite restrictions.)
3b/ 2. Using detailed census data from 2016, what neighbourhood characteristics/factors correlate (or do not correlate) with % pos / cases & their recent weekly trends (since Aug 30). (Characteristics include industry employment, % low-income prevalence, and % visible minority)
4a/ Data note/sources: - Toronto began publishing weekly neighbourhood test % positivity and tests per 1,000 (weekly), and from population data, we can infer cases per 100,000, and their trends.https://www.toronto.ca/home/covid-19/covid-19-latest-city-of-toronto-news/covid-19-status-of-cases-in-toronto/ …
4b/ Data note/sources:
- for completeness, below are the neighbourhood dataset limitations/caveats from the Toronto #COVID19 tracker website. Note % pos & cases understated, but *directionally correct*.pic.twitter.com/NcPzuAcXOq
4c/ Data note/sources: Link to the neighb'hood census data used: https://open.toronto.ca/dataset/neighbourhood-profiles/ …pic.twitter.com/WAtnl4jUFT
6/ Here is the overall picture in Toronto since August 30th. Overall rising cases and positivity. (Toronto population ~2.73m people (per 2016 census))pic.twitter.com/IUN0jRkee5
7/ But what proportion of the city is experiencing an increase vs. a decrease in those metrics? (defining a decreasing neighbourhood, somewhat arbitrarily, as one that experienced a two week in a row decline in % positivity AND cases per 100,000)
8/ 97 neighbourhoods, representing 2.0m people (73% of the pop.), are experiencing rising positivity and cases per 100,000, and account for 84% of new cases in the latest three weeks of this dataset (ending Oct 31st).pic.twitter.com/BWuXX71kMj
9/ Conversely, the other 43 neighbourhoods, representing 750k people (27% of the pop.), are experiencing flat/declining positivity and cases per 100,000, and account for just 16% of new cases in the latest three weeks.pic.twitter.com/flGI55cPi8
10/ What if we cherry-picked a couple of example areas? Like the Downtown Core (“DTCore”)? How is it doing? Here, %pos & cases/100k peaked at the end of *September*, well in advance of the Oct 10th restrictions. These 10 hoods have a total population of 300,000 (11% of the pop.)pic.twitter.com/1UOoAVvvPl
11/ Northwest Toronto (“NWT”), on the other hand, is fairing worse. Some of the highest positivity rates and continued increases in % positivity (now >9%), despite the Oct 10th restrictions. 282k people (10% of pop., and 22% of recent new cases).pic.twitter.com/Ro0R7KMiwm
12/ What’s driving the differences? Detail later, but for now just note the following: % of workforce in service industry (defined later): NWT-72% DTCore-41% % of workforce in knowledge/work-from-home industries: NWT-19% DTCore-50% Average Income NWT-$28k DTCore-$49k
13/ So obviously we need to explore some of the neighbourhood specific socioeconomic/demographic factors and see how/if they correlate (not necessarily “causate”) to %positivity and case trends. I will focus mostly on workforce composition.pic.twitter.com/ljFn9wTk0l
14a/ Workforce composition. The Toronto census data recorded # of people in 20 industry groups (see picture below). I categorize the 20 into four “groups”: http://1.Services 2.Knowledge/Work-From-Home (“WFH”) http://3.Education http://4.Healthcare pic.twitter.com/H35xe3Htze
14b/ Note that I subdivide Service into further “subgroups” into (i) Retail & Other, (ii) Foodservice & Accommodation, (iii) All Other Services. See picture again. Industry legend source: https://www23.statcan.gc.ca/imdb/p3VD.pl?Function=getVD&TVD=118464 …pic.twitter.com/DhxAgXgC2F
15/ How does the % of a neighbourhood’s workforce in a given industry “group” (i.e. Service or WFH) correlate with total positivity, maximum positivity, and cases per 100,000? Well, here’s how (shown in table form here, and graphic form to follow):pic.twitter.com/GaqaNy8PJL
17/ We see Service (i.e. community) type jobs strongly positively correlate to % positivity and total cases and knowledge/work-from-home jobs strongly negatively correlate to positivity and cases. Perhaps it is no surprise, but the correlation levels/consistency is incredible.pic.twitter.com/cqigyZCze3
19/ Let’s visually examine *trends* in positivity for neighbourhoods with the top 25 & 50 highest % and 25 & 50 lowest % of their workforces in either Service or Work-From-Home Industries, in this recent current “wave” (e.g. since August 30).
20/ This chart is incredible. The average positivity trend for the top 25 & 50 neighborhoods by % of workforce in the Services industries is a straight line up. And neighbourhoods with the lowest concentration of Services workers…positivity barely budges.pic.twitter.com/NkCA4d3FTR
21/ Looking at it the opposite way... i.e. Total % positivity of the neighbourhoods with the 25 & 50 highest vs. 25 & 50 lowest concentrations of knowledge/WFH workers. Neighbourhood % positivity barely budges for the highest concentrations of these workers.pic.twitter.com/gcjkdb68a5
22/ So what about those pesky bars and restaurants? With positivity correlating with Services workforces, doesn’t that mean bars and restaurants are the big culprit?pic.twitter.com/GHmb29gKQv
23/ Well, it appears somewhat. BUT, of the three Services industry subgroups (foodservice/accom, retail & other, all other services), neighborhood concentration of foodservice/accom industry workers is the *LEAST* positively related to increasing % positivity:pic.twitter.com/Zb2494mcmI
24/ Quickly touching on % visible minorities and non-visible minorities.
Clearly visible minorities are struggling with #COVID19. The 50 neighbourhoods with the highest visible minority concentration account for 41% of the pop., but have 53% of cases since Aug 30, w/rising % pospic.twitter.com/nlBB5Ix13r
25/ % visible minorities in a neighbourhood correlates with concentration of service workers in a neighbourhood, so as we have seen above, occupation could be the driving factor. I will leave it to the immunologists/virologists to comment on immune susceptibility differences.
27/ So it appears the task & health cost of containing #COVID19 thru lockdowns/societal restrictions *FALLS HEAVIEST* on lower-income neighbourhoods, with minorities and service workers, who *must* go out into society/community to earn a living, and who may be more susceptible.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
