We have already had a lot of COVID-19 articles. Every country's goal now is how to reduce the number of death. Since the exponential growth rate of an epidemic, it should be the first priority that how to avoid the corruption of the health system.
We need the precise data of how many patients are there. Otherwise we can not have a strategy, how many beds we need, how long we need to lock down, how many test we need to deploy, every strategy needs to be answered the questions: how many X we need?
In this article I based on the data from John Hopkins Data Repository and the analysis of Aatish's Exponential/Logistic Curve-Fitting Site.
When I checked this data, I found an abnormality of Japanese data.
Figure 1 shows the total confirm cases and new confirmed cases in the past week. Japanese data is way less compare to other countries.
Figure 1 Data source: minutephysics: How To Tell If We're Beating COVID-19
You might think, a small new case ratio sounds good. However, why Japanese case is small? It is the same disease. Some countries completely lock down a whole city, but even so, this trend kept. There is no complete lock down in Japan yet. This figure is log-log scale, the different is looks small, but it is a few ten times.
The next Figures are about the speed of spreading (doubling time). Only three countries including Japan is over the ten days. If the error bar is concerned, only Japan is quite certainly over ten days. Note: these two graphs have a different Y axis due to automated graph generation. For an exponential growth, one day difference affects big.
Figure 2: COVID-19 nation vs. doubling time (1) Including Japan (only a few over 10 days)
Figure 3: COVID-19 nation vs. doubling time (2) (no over ten countries)
Figure 2, 3 Data source: John Hopkins Data Repository https://github.com/CSSEGISandData/COVID-19 and Aatish's Exponential/Logistic Curve-Fitting Site https://github.com/aatishb/covid/blob/master/curvefit.ipynb
I am not an epidemiologist, thus, I cannot tell, but I have a question, why? This looks not normal. What is going on?
We have a few proven effective methods for COVID-19,
Some of them can be done by individual. But tests are one of society's work. If we know more, we can possibly do more. If we don't know we can do less. Some countries achieved to deploy the good test system.
We can have a strategy based on test results information. How many beds we need more in the next week, how long we need stay home, these strategy has "how many/how long" questions. The data is necessary for saving lives. But I see strangeness in this data. That is my concern. Although this is my personal observation. How do you think?
I hope this is just my concern.
I found the following links useful to understand the data.
We need the precise data of how many patients are there. Otherwise we can not have a strategy, how many beds we need, how long we need to lock down, how many test we need to deploy, every strategy needs to be answered the questions: how many X we need?
In this article I based on the data from John Hopkins Data Repository and the analysis of Aatish's Exponential/Logistic Curve-Fitting Site.
When I checked this data, I found an abnormality of Japanese data.
Figure 1 shows the total confirm cases and new confirmed cases in the past week. Japanese data is way less compare to other countries.
Figure 1 Data source: minutephysics: How To Tell If We're Beating COVID-19
You might think, a small new case ratio sounds good. However, why Japanese case is small? It is the same disease. Some countries completely lock down a whole city, but even so, this trend kept. There is no complete lock down in Japan yet. This figure is log-log scale, the different is looks small, but it is a few ten times.
The next Figures are about the speed of spreading (doubling time). Only three countries including Japan is over the ten days. If the error bar is concerned, only Japan is quite certainly over ten days. Note: these two graphs have a different Y axis due to automated graph generation. For an exponential growth, one day difference affects big.
Figure 2: COVID-19 nation vs. doubling time (1) Including Japan (only a few over 10 days)
Figure 3: COVID-19 nation vs. doubling time (2) (no over ten countries)
Figure 2, 3 Data source: John Hopkins Data Repository https://github.com/CSSEGISandData/COVID-19 and Aatish's Exponential/Logistic Curve-Fitting Site https://github.com/aatishb/covid/blob/master/curvefit.ipynb
I am not an epidemiologist, thus, I cannot tell, but I have a question, why? This looks not normal. What is going on?
We have a few proven effective methods for COVID-19,
- Take a social distance
- Hygiene habit
- Test and quarantine
- Better healthcare system
Some of them can be done by individual. But tests are one of society's work. If we know more, we can possibly do more. If we don't know we can do less. Some countries achieved to deploy the good test system.
We can have a strategy based on test results information. How many beds we need more in the next week, how long we need stay home, these strategy has "how many/how long" questions. The data is necessary for saving lives. But I see strangeness in this data. That is my concern. Although this is my personal observation. How do you think?
I hope this is just my concern.
I found the following links useful to understand the data.
- Exponential growth and epidemics (3Brawn1Blue)
- The Coronavirus Curve - Numberphile
- Simulating an epidemic (3Brawn1Blue)
- How we must respond to the coronavirus pandemic (TED)
Comments