6 ms·
Evidence of synthetic data in Wuhan coronavirus body count
- anotheryou 7y agodamn, I was really worried about the hockey stick on the infection rate and it became more linear, but now I guess I can't trust any numbers...
- SpicyLemonZest 7y agoThe reported count for yesterday is already 3000 off the trendline. I suspect it’s easy to draw a quadratic through 15 data points for any uncontained infection, since the first and second derivatives are sure to be positive.
- anotheryou 7y agoalso death counts? prediction: 808 reported: 811 -- sources: EN: https://www.wsj.com/articles/coronavirus-outbreak-has-killed-more-people-than-sars-11581226253 https://www.wsj.com/articles/coronavirus-outbreak-has-killed... probably the original: http://www.nhc.gov.cn/yjb/s7860/202002/4f28ab5ca87d42d284833df3ccc8d45a.shtml http://www.nhc.gov.cn/yjb/s7860/202002/4f28ab5ca87d42d284833...
- SpicyLemonZest 7y agoThis is data for 2/8, where the redditor’s death count prediction was 721.
- ryanwaggoner 7y agoI think their prediction is for report date. Between reporting lag and time zone differences, I wouldn’t discount this if it’s tracking the trend lines but just off by a day.
- anotheryou 7y ago2/8 at 12pm, published 2/9 I aggregated the numbers for fatalities: date published | reported (count as of 24:00 the previous day) | predicted 2/5 | 490 | 489 2/6 | 563 | 561 2/7 | 636 | 639 2/8 | 722 | 721 2/9 | 811 | 808 2/10 | ___ | 900 2/11 | ___ | 997 so far +/- 3 difference
- anotheryou 7y agodate published | reported (count as of 24:00 the previous day) | predicted 2/5 | 490 | 489 2/6 | 563 | 561 2/7 | 636 | 639 2/8 | 722 | 721 2/9 | 811 | 808 2/10 | 908 | 900 2/11 | ___ | 997
- anotheryou 7y agodate published | reported (count as of 24:00 the previous day) | predicted 2/5 | 490 | 489 2/6 | 563 | 561 2/7 | 636 | 639 2/8 | 722 | 721 2/9 | 811 | 808 2/10 | 908 | 900 2/11 |1016 | 997
- ryanwaggoner 7y agoNot sure what you’re looking at. Cases for Feb 9th is 37,251 and Redditor prediction was 38,675. Yesterday was even closer: 34,506 predicted vs 34,598 actual. Source: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports https://www.who.int/emergencies/diseases/novel-coronavirus-2...
- tim333 7y agoI believe the Chinese numbers are supposed to be for people tested and found to have the coronavirus and for the people out of those who died. This means they are limited by the number of tests carried out which may explain why the numbers go up in a fairly smooth way. An exponential boom in actual infections would just mean most of those would be untested. So rather than synthetic data maybe actual data but limited testing.