3 ms·

The numbers from Wuhan are simply wrong and we have no idea how many people there are infected or how many have died. So you should discard all data from Wuhan

by mantap 7y ago

The numbers from Wuhan are simply wrong and we have no idea how many people there are infected or how many have died. So you should discard all data from Wuhan at least before doing your analysis. The whole city is under quarantine, people may not be able to get to a hospital. Even if they do, hospitals are overloaded and beyond capacity. If people there die before being tested their death won't be counted as being from coronavirus. If they have mild symptoms but are not in a serious condition they would be stupid to go to a hospital and get tested (and risk getting infected). And China is a communist country so the numbers should be taken with a large grain of salt anyway.

viraptor 7y ago
People have really strong reaction here. Outside of popular news, we know how to deal with incomplete data and biased numbers. This probably should be told more often when reporting the cases on the TV, but I expect anyone actually having impact on the issue understands that any actual counts we get are a lower bound, that testing resources and travel regulation affects the reports, and how to take into account other reporting biases. This is not a new idea.
- mantap 7y ago
  I was replying to this: "I just copied the data from WHO sitreps and the total-cases diff of diffs has been increasingly negative for three of the last four days." In any case I disagree that the data from Wuhan contains any information. Even if you can magically model away all the problems in the data, the Chinese government is still a disreputable source. Chinese state media is notorious for only showing the CCP in a good light. When you see the words 'Xinhua News', what comes to mind? Propaganda right? I don't understand why people think that China is publishing the truth now, when China is in crisis and freedom of speech is trending on Weibo. Surely this is the time to turn propaganda up to 11. These numbers are the most powerful tool that the Chinese government has for propaganda right now. And given their penchant for it, the baseline assumption should be that they are being massaged into the narrative that Beijing wants to tell.
- PeterisP 7y ago
  We know how to deal with random noise in the data and random omissions in the data. However, if data is completly fabricated with no input from reality, then it's not possible to tell anything about reality from that data, only parameters of the (fictional) model used to generate that data. We can determine the assumptions about the disease which were encoded in that model, we can't validate whether these assumptions match reality.

by mantap 7y ago

viraptor 7y ago

People have really strong reaction here. Outside of popular news, we know how to deal with incomplete data and biased numbers. This probably should be told more often when reporting the cases on the TV, but I expect anyone actually having impact on the issue understands that any actual counts we get are a lower bound, that testing resources and travel regulation affects the reports, and how to take into account other reporting biases. This is not a new idea.

mantap 7y ago
I was replying to this: "I just copied the data from WHO sitreps and the total-cases diff of diffs has been increasingly negative for three of the last four days." In any case I disagree that the data from Wuhan contains any information. Even if you can magically model away all the problems in the data, the Chinese government is still a disreputable source. Chinese state media is notorious for only showing the CCP in a good light. When you see the words 'Xinhua News', what comes to mind? Propaganda right? I don't understand why people think that China is publishing the truth now, when China is in crisis and freedom of speech is trending on Weibo. Surely this is the time to turn propaganda up to 11. These numbers are the most powerful tool that the Chinese government has for propaganda right now. And given their penchant for it, the baseline assumption should be that they are being massaged into the narrative that Beijing wants to tell.
PeterisP 7y ago
We know how to deal with random noise in the data and random omissions in the data. However, if data is completly fabricated with no input from reality, then it's not possible to tell anything about reality from that data, only parameters of the (fictional) model used to generate that data. We can determine the assumptions about the disease which were encoded in that model, we can't validate whether these assumptions match reality.