Ho-hum… We remain locked-down… So not much news to report I am afraid, no adventures, no brushes with excitement, nothing. But life does continue. For now the infections as a percentage of tests remains fairly constant, so whilst the number of cases of Covid-19 continues to grow, it is not exactly ‘exploding’, yet. I think this trend is best illustrated by these two graphs:
Having said that though, there are two areas of concerns for me. The first area of concern for me is how many days it takes to double the total number of infections. I have been looking at the infections per day (starting from 11 March) and counting how many days it took to double the number of reported infections. At the beginning (with a small total number of infections, the doubling happened every two or three days. Then just after two weeks into the story, the doubling suddenly jumped to 14 and then 18 days. This was a very good sign. It was hard to attribute, but it was possible that this was driven by the fact that 10 days previously we had started social distancing measures. Unfortunately after stabilising for a while at taking 17 or 18 days to double the infections, from 1 April it started taking less days for the number of infections to double. It took just 15 days for the number of infections on 3 April to double and I suspect this number may continue to fall over the next week.
The increasing mortality rate (as a percentage of infections) is also an area of concern. I have been tracking two metrics here. Firstly deaths as a percentage of infections recorded on that same day and secondly deaths as a percentage of infections recorded two weeks prior. The thinking here is that as it takes approximately two weeks for people to become symptomatic and thus to become ill enough to present at hospital before succumbing to the illness, mortality is a significant lag measure and to understand true mortality rates at any given time, you need to look at how many people were infected at least 14 days ago. As can be seen from the graph mortality rates are steadily rising and on the latter metric are sitting at 3.5% which is very close to the 5% reported by countries who were experiencing significant strain on their medical services. This is of great concern as according to our department of health our hospitals and ICUs are not yet under much pressure at all. As the infections increase and as more people are admitted to hospital this mortality rate will probably increase, perhaps dramatically.
These graphs have resulted in some interesting ‘push back’ on Twitter. I decided to track the data coming out of the NCID for my own purposes so that I could perhaps better understand the situation and as I like seeing what can be gleaned from the numbers available. Note all the information here is gleaned from just three metrics published daily by the NCID: a running total of the number of tests conducted, a running total of the number of infections recorded, and a running total of the number of deaths recorded per day. Yes, I am a geek, I enjoy ‘manipulating’ data to see what can be learned. I am an amateur and by no means a professional statistician. I know my limits and thus do not run any statistical models etc. I simply present the data as is and sometimes provide a little commentary/opinion. Having done the work, I thought ‘why not share it on my Twitter feed?’ My thinking was that if people found it interesting/helpful/useful, great. If not they could ignore it. I was doing the work anyway, I may as well share it. I felt ‘safe’ doing so as the data was literally just the official NICD data repackaged in graphical format, with no extrapolations or predictions.
Anyway, I had been merrily posting for a number of days and making far fewer observations/inputs than above, when someone who follows me told me that he felt I was wasting my time doing the graphs daily and that I should confine myself to weekly numbers.
Now this is arguably a valid point, a 0.1% shift from Monday to Tuesday is probably not significant or meaningful. However for me doing the work daily is less onerous (inputting a few numbers into a spreadsheet literally takes a few minutes). If I were to do this weekly it would take much longer to collate the data, so for me, it is better to do the work daily. I also think that if I only published the graphs weekly fewer people may actually see them (a lot gets lost in the Twitter noise), so whilst some people may see the graphs every day, others will only see them every four or five days and that is also ok. You can still see the graph and make whatever sense of it you will. So I was happy to ignore this criticism.
However he followed up by seeing that it was useless to add trendlines as the graphs showed nothing. Now this was annoying. None of my graphs have trendlines. I have deliberately tried to avoid adding much by way of analysis or developing theories around what will or will not happen. All that I did (usually in response to a comment or query, whether online or in person) was maybe say that looking at the data already in the system (eg with respect to how many days before the infections would double) that we can probably expect the number to increase or decrease. What I was prepared to say was whether a development was positive or not. But I deliberately tried to refrain from making predictions and assumptions outside of the data that was readily available. I also refused to engage in speculation as to whether the underlying data was correct, reliable or not. I felt that this was well beyond me. I therefore felt somewhat aggrieved that I was being accused of doing something I had not done. But anyway…
Later another person who follows me said that he thought my graphs were complete nonsense. That they did not say what I thought they said and that I was doing a thoroughly bad job of labeling the graphs. I note he did not say specifically what could be improved or what was wrong with the graphs, just vague criticisms especially of what I allegedly had said the graphs proved/showed, when I had specifically avoided saying much of what the graphs said beyond describing an ‘up’ or ‘down’ tick and whether that was ‘good’ or ‘bad’. So I would say an uptick in infections was ‘bad’ whilst an uptick in the number of days to double the number of infections was ‘good’. I am not sure how you can take issue with these statements.
So two men who I did not know from a bar of soap, who were neither my employer nor family members felt they could dictate to me what to do with my time and critique in very broad unhelpful (to me or anyone else reading) terms my work. I was less stung by the criticism (which I was happy to ignore, especially as others on other platforms who were actual specialists in this field had taken no issue with what I was doing) as I was stung by the attitude displayed by these two people. In fact had anyone provided constructive criticism I would have happily taken it. Had someone said, good work, but you could improve the graph by adding this or that. Or indeed if he had said I am not sure about this, perhaps if you labelled it differently I could make better sense of it, I would have happily accepted this feedback. But this is not what was said. At all.
Now what is especially interesting to me is that I had been sharing these exact same graphs with similar observations and comments on my ‘boy’ Facebook page. Here I had actually received positive feedback from actual medical (and other) researchers (from various parts of the world). Some people, especially I suspect those from overseas, had even started following me for the graphs as they gave them ready insight into the South African situation, which they were interested in, but as you can imagine a little too busy to deal with directly. Not one of these people found fault with what I was doing.
Yet on my ‘girl’ Twitter account men who (as far as I know) are not medical researchers, felt compelled to criticise and attempt to silence me: to be clear telling someone to stop doing what they are doing, to stop publishing what they are publishing and that their attempts are worthless is a deliberate attempt to silence that person. The complete distinction between the reactions is dumbfounding… Until you realise that this is in fact common place. When presenting as a woman I get treated as a woman. That means I get to experience everything women get to experience. Both the vaguely positive (men helpfully assisting me with shopping trolleys in the mall parking lot) to the definitely negative which can range from men making lewd comments in the street, to being silenced for having the temerity to intrude on a space (the world of data, analytics and GRAPHS) that some men seem to think is their domain.
Perhaps I should just retreat to the kitchen and make some banana bread, or whatever the lockdown baked item du jour is currently.
The gender differentiation thing can become very obvious; I see it all too often on social media.
Your graphs are pretty interesting; the days to double is also one not seen much. I remember a month ago, infections in Italy etc. were doubling ever 2-3 days, and that was considered terrifying. I think we’re doing reasonably well. I’m not a big fan of infection numbers because they seen to depend a lot on how many tests are being done, and where. Death rates seem a little more concrete to me. I’m no expert at all though, I just look at figures going around. I’ll let the experts draw the conclusions.
LikeLiked by 1 person
Another great post.
It’s almost like you are a walking human psychology experiment gauging how gender affects the way we may be treated by others.
LikeLiked by 1 person
lol… I am a walking experiment in many many ways. 🤣🤣🤣
LikeLiked by 1 person
Thanks for the kind feedback
LikeLiked by 1 person