“Data Science with Open Canada Data” Weekly Seminars
www.IVIM.ca / data / seminar
29 April 2022: Update on Cases following vaccination, Excess Deaths in Canada, Death tracker app
Dmitry Gorodnichy, dg@ivim.ca
Transcript
Hello everyone, welcome to our seminars on the analysis of official canada data related to covid19 vaccinations.
my name is dmitry gorodnichy. i'm a data scientist with over 20 years of experience in data science. so this is your opportunity to ask as many questions as you wish related to
official government canada data related to vaccines.
i know that vaccines, specifically covid19 vaccines, have become a very debated polarized topic in our society, and we hope that data can help different populations in our country to talk about this subject, because data again is just data. it doesn't support conservatives or liberals. it doesn't support truckers versus non-truckers , it is just data, and it sometimes supports one group
sometimes it supports the other group. and the best way to see when it supports which group is by visualizing these data, because data also changes over time, and this is very important to appreciate.
we live in a changing world. in our life in general, nothing is stable, nothing is static. things are changing over time and this relates to vaccine efficacy, vaccine safety.
it relates to deaths. and yes there are many ways to look at this data, and here we show you the tools which would allow you to see the data better .
it's the same data which is published by official government canada websites. we just take this data and we visualize it. so today specifically we'll be talking about the excess deaths. but before talking about that let me remind you where you would find all the information about the analysis which we conduct. it's shown on your screen right now. we have a web page dedicated to the data seminar or webinar. we have it every friday at noon Ottawa time. and anyone can join and ask questions via chat or later offline on youtube, because we post all these recordings on youtube. and we would every friday would discuss a particular topic which is of interest to many of us.
so this week we're talking about excess deaths. specifically in the year 2020 2021, and we will also provide a little update on cases following a vaccination. that's what we always do in the beginning of our seminars. so, let's jump right into the latest update. for this we go to the main dashboard, which is shown on your screen, and i suggest we just scroll down to the very bottom of the screen, which shows the table .
this table is nothing but numbers which have been pasted from the official government canada website related to cases following the vaccination. and when we look in this blue portion, these are the numbers which are copied directly from the source, as you would see from the source, which are pasted right here. you can go yourself there or you could just look at the numbers pasted here. what do we see here ? we see that the numbers are going up again, up and up again. by about 0.8 percent since last week the percentage of death among full vaccinated has increased, right ?
so, and we have graphs which plot exactly the same information but using just more convenient data science tools. so these graphs show exactly the percentage of fully vaccinated among cases - infections, hospitalizations and deaths. so what do we see here? the last report as of today was the 10th of april - and you could hover with your mouse here and it shows you there the exact numbers here, as reported by public health agency of canada. and we see that indeed the blue bar, which is infection, have increased from 45.9 to 46.6 counting cases since december 2020. now we take this data and then we compute number of cases in the past week, in the past week. and then we compute the percentage of full vaccinated among those cases in the past week. and this is shown right here so what do we see here we see number of deaths 75 %. this again is percentage of fully vaccinated among deaths, as opposed to smaller number which was 70 percent last week. so we see the dynamics is shown to you right here on the screen. so these are very convenient toolls and graphs which allow you to easily visualize the data which are plotted by public agents of canada website, and again the website is right here it's this link which is shown right here it's epidemiology summary covered 19 cases you can go there and it's table number two and we just plot this number. so that was quick update on the cases following vaccinations. today our main subject is excess deaths, and you would see that on our main dashboard - which is right here on the portal ivim.ca - we have added this new graph, a new block, dedicated to excess deaths .
and what it shows you ? it shows you by province the number of deaths per week, in particular, for the age group from 0 to 44 years of age. and again as always, today we take official data. in this case it's from statcanada - statistics canada website . and you can click on this link and it will lead you to this website - which I'm showing you right here - which is called provisional weekly death count by selected group causes of deaths. and you would see, this website provide you options to download the data. you can click on download button and you can download this csv file. CSV stands for comma separated values. it's just a file which contains lots, lots of numbers. and we download this file and we plot it, simply plot it using a nice data graphics package which allows us to plot the total number of deaths by week for each province. so let's have a look what we are seeing here ? we see that in alberta the red line, which is the number of deaths per week since 2020, is going slightly up. again, just look at the horizontal line - it shows you the dates starting from the january 2020. that is the beginning of pandemic. you see that indeed since the beginning of pandemic we have excess deaths - about maybe 10-15 deaths per week in alberta, right ? then we started vaccination, and the gray line here shows you 1 january 2021. and you see we still have excess deaths here.
now, the thick gray line shows you when the vaccination has been extended to general population. you remember, in the beginning it was only a high risk population who was vaccinated. and then in about the first of april - which is what is shown by this thick line - we have vaccinations extended to the entire population, to otherwise healthy canadians. and that's what we are looking here. what happened after the vaccination has started with general public ? we would of course hope that vaccination would help to reduce the total number of deaths. that's the key objective of any procedure or any measure which is undertaken by or in any country: to do something - whether it's quaranteening, whether it's lockdown, whether it's vaccination - to do something which would reduce the mortality, mortality rate, right ? and everything the government is doing is really trying to help to reduce the mortality rate, because, yes we see the mortality has increased in alberta, in british columbia, in canada national, in manitoba, you see it was
higher than normal. so what is normal here? in this plot, blue line is the average two years average before pandemic so this is 2018 - 2019 average data for each province total deaths.
now what is interesting that in new brunswick you don't see much excess deaths during the pandemic but again you still see in new foundland, in nova scotia, in ontario, right ? and in quebec and saskatchewan, yeah? now okay we had excess deaths - more people are dying than normal during the pandemic. the governments in province have their own measures or own policies - and you know the policies in alberta, for example, are very different from quebec right now or ontario. and in saskatchewan, for example, they are no longer even a measure cases, covid cases from ,what i understand. so we have different policies in different provinces, and also we have potentially different vaccines administered in different provinces.
so and maybe not only a type of vaccine, maybe also different batches, different lots of vaccines, and recently there was a news that moderna recalled one of their batches because it was contaminated, for example - i put it on my twitter here that indeed it happens sometimes. (https://www.reuters.com/business/healthcare-pharmaceuticals/moderna-recalls-thousands-covid-vaccine-doses-2022-04-08/ - Moderna recalls thousands of COVID vaccine doses in Europe, April 8, 2022 (posted https://twitter.com/Gorodnichy_Dm/status/1512783260367134727)
it happens in food industries - sometimes food is contaminated and there are recalls, because people are getting sick and poisoning from food infections, right? and the same happens, of course, in any industry, and i guess in vaccination. It's not an exception. Vaccination is not exempt from this situation, and this potentially could explain why we have such differences, why we observe such differences in the excess deaths in different provinces, because here we do see huge difference between atlantic, in particular new brunswick, and let's say quebec. let's start from quebec .you see in quebec, young people or only 44 years of age they have less mortality rate after vaccination than before vaccination. this is excellent news for quebec, assuming of course that the data is recorded properly and we trust the data, which we do. I believe the data is the best source of trusted data, which is data published by a government of canada agency, is the most trusted unbiased piece of information which all of us, regardless of our political or religious beliefs, can rely on. so that's what we see in quebec .
now what is happening in new brunswick ? here's definitely something started happening right after the vaccination. again we don't know what. this could be related to some other health issues, maybe there was another pandemic there, maybe there is poisoning there, maybe there is a wave of suicide there, we don't know. what we see is only that the death rate has increased from about five deaths per week in historical average to 10, then 15 then 20 now to 25. now this drop here - just ignore it, it needs to be ignored, because it is just because many provinces, they just put zeros when they don't have data, when they have no data yet they put zeros. and that's why it shows as zeros. there are zero deaths here, and there is a steep decline because we take the average of three weeks, just to smooth the data we take the average the running average for three weeks. so just ignore the data where where you see the curve goes down. so what do we see here is really we see it going up up and in fact, if we put a trendline, really we would like to see what is happening then? is it going to continue going up ? hopefully not. i hope it will go down.
but let's see what happens in other provinces ? so you see in alberta also it started to go up after a general public vaccination . the same in british columbia it started going up after general republic of population, manitoba stopped reporting deaths altogether since about june last year, so we just can ignore it. now new foundland = also increased by actually by more than 100 percent in excess deaths after vaccination, nova scotia - nova scotia not as much but still some, ontario - also you see that excess deaths after vaccination is more than excess deaths before vaccination even during the highest peaks, or waves, of pandemic.
now we don't know what is causing these deaths. what we do, what we do know - is that currently canada, as the entire globe, is experiencing the wave of omicron variant, and there are many sources right now - my favorite is john campbell from united kingdom, who every day provides some insights from again from official data - it's very important that he's like myself using only official data here - which shows that
right now 99 % of all cases is omicron, and it's apparently - or not apparently, based on the data -
it's a very safe variant, meaning that most people who have it, majority like 99%,
don't have complications from omicron.
but it's very infectious, so it's extremely infectious, you cannot stop it. in the united kingdom it's more than 90 percent are now essentially already infected by omicron. so everyone in hospitals, everyone who is dying right now, if they have been tested for it, if they had the test, most likely this test would show 90 percent likelihood that it has omicron, and these people would still die if they were unfortunately having diseases. so you would see that number of hospitalizations with omicron specifically with omicron, and number of deaths with omicron, of course, jumps significantly. now what the people are dying, what they are dying from ? - again based on the fact that omicron is much safer, much less immortal or lethal, compared to delta and the original wuhan variant. omicron is the safest
not the safest i shouldn't say, but 'the safest' meaning
it has much less mortality rate than previous variants. we know that. so based on that we can assume that there are less deaths of omicron now compared to the number of deaths of delta or wuhan original or alpha variant right? so based on that, we would understand that this increase in deaths is very unlikely because of omicron,
being impacted by omicron. it's something else and we don't know what, and here we're just data ,
we're data scientists here. we do not make any conclusions. we can ask questions and we do ask questions. we can make observations, and we do make observations here. but anyone can make this every observation yourself, right ? and please use the chat window right now here or if you're watching this video offline, please ask questions directly on youtube - we will make sure we will respond to all your questions here.
please don't forget to subscribe to this channel. and also please also subscribe us now on a facebook page. we have just created last week we have created a facebook page which is shown right here it's called open open canada data. and uwe now have 38 members in one week and we also have twitter i have my personal twitter here, which now has 125 followers -
just one, one who joined us right now as we speak, which jumped from 60 last week. so twice as many. again we just try to
talk about data and to help people to see the data better using contemporary techniques. now what i'd like you to show you - how you can visualize this data yourself. this is done with the death tracker app. and you would see here - we have a collection of applications built specifically to help you to visualize various data coming from official sources. now if you click on the tracker, you will get to the to this app, which is interactive and which reads data directly from statistics canada website
provisional weekly discount by causes, and it would allow you to play with it. it's really ...
it's almost like a game. it's interactive and you can invite your children, your neighbors, and play with it. now, you have to to allow the program to load the data - it takes about one minute or half a minute - and you would see the error message (you just saw the error message )
when it loads the data, when there is no data, it shows the error message. just ignore it. you have to wait for about maybe 30 seconds, because again it's getting the data directly from the source .
and it shows you this graph now.
this graph shows a little more than what i've shown you there, because here it shows you by causes. this app has been built about four months ago and it allows you to track causes of deaths. how you do it ? - you can select here which causes of deaths do you want to plot and ,
let's say, right now i'll just select total causes and cancer, which is neoplasm. okay, so you have one row, top row shows you for total causes and the other one shows you for cancer, and we have added a new functionality here. which was not there last week. it's called show excess death. and that's how you get this black line added on top of the other line. so let's see what happens when you click here ? when you click here, when you don't have this functionality ,
it will just show you the changes of deaths over time. and it's very interesting information because you can even see the dynamics for example for flue, for example you can add also blue and the way you do it you just hold ctrl key and click on this to add you can add also suicide which is intentional self-harm. so you see we now have four rows here. and then you could do some analysis of this. for example you could show a trendline .
you could just put a line and see whether it goes up or down . and here you would see for example that for third line. which is in influenza, essentially uh we have much less influence cases during the pandemic and people know that's the case but also very interesting we have less cases of suicide right? and this, of course, becomes a very interesting observation,
because in fact during the pandemic it was known that people were very much distressed being laid off, being put on live without pay, not allowed to travel, to attend social activities, sports activities, cultural activities for example, if they are not vaccinated. so it would have been expected that accidents,
i mean suicides not accidents sorry ,
suicides would go up. but they go down. so this is indication that maybe, maybe some of those suicides, they were labeled as covid cases,
because they were also maybe having covid
infection.- and i just have selected covid deaths here.
now when you click on 'show excess deaths',
it will just compare theses curves to historical average. and now you can compare what you have now in 2022 or 2021 with a historical average shown by a redline. so that's pretty much it.now you can also ask the application to show you the `absolute difference `,
and it just subtracts the numbers. the black line would show you zero, meaning there is no change. if it's above black line, it means you have excess deaths, and this is where we would see that we have excess deaths in all provinces, maybe not as much in quebec right now. or you could also show it as a `relative difference` as a percentage,. by how much we have this? and this is where you see by how much it's by for new brunswick for example it really goes by more than 25 here, by 30 up so it gives you numbers to quantify the excess deaths before
vaccination and after vaccination. and these dash lines vertical dash lines show you the beginning of pandemic the first dash line, then the second dash line shows you the
is before the second wave, and this is also when the policies have changed in some provinces .
then the third dashed line shows you the beginning of vaccination , and then the last dashed line shows you the vaccination with the general population. so you can play with it, you can show it using different scales. for example logarithmic scale can be used sometimes it's very convenient it shows you how much increase you could also add for example vaccination rate you can show ask to show vaccination rates if you don't remember what it is but most people already know it you know how it was going up it was started it started from zero in january 2021 and it reached 80 in about november december but then you can really you can compare and you could correlate these curves with the other curves here. so you can really have a good time exploring this app. and we will be adding more functionalities there soon.
so that's pretty much what we wanted to say. thank you everyone for watching .
and again please join us on ...
this is the channel - independently verified information , on facebook group` open candidate data`, on twitter Dmitry Gorodnichy,
and keep following us on our main website ivim.ca.
next seminar will be held as always on friday next week and we would normally announce it here on this web page : / data /seminar .
and we will also post slides, transcripts, and announcements right here.
thank you everyone everyone for watching, be safe and enjoy. spring is coming.
Appendix:
The results for excess deaths in year 2020 and 2021 are in. They look nothing but normal, and very different from one province to another (See image below. Images for various age groups are here).
What do they tell us about pandemic and vaccine safety and effiacy? We will discuss it this week!
Join us this Wednesday noon via Facebook event, or this Friday noon via Zoom link below as usual.