Weekly COVID-19 (Coronavirus) statistics and analysis

See our Coronavirus data Dashboard for individual country data.

COVID-19 statistics and cases around the World


We think there is some effect from seasons, but poor or delayed responses swamp any benefit. In wealthier southern hemisphere countries, like Australia and New Zealand, the cold weather does not appear to have had an impact on the spread.

Latin America

Within Central and South America the virus is now spreading at a substantial rate with some countries reporting double and triple numbers compared to last month.


Second Wave


More COVID-19 Statistics and Analysis

See our latest blog posts or podcasts here. See our Coronavirus data Dashboard for individual country data

Data sources

This is a list of some of the main data sources we use:

https://www.worldometers.info/coronavirus/ Probably the best source of the latest COVID-19 statistics

https://medium.com/@tomaspueyo/coronavirus-the-hammer-and-the-dance-be9337092b56 Tomas Pueyo has written a number of very good summaries of the strategies to overcome coronavirus

https://www.capitaleconomics.com/the-economic-effects-of-the-coronavirus/ Good source of fast-moving China economic stats.

https://bnonews.com/index.php/2020/02/the-latest-coronavirus-cases/ If you want to be bombarded with every breaking news story, this is the place

https://ncov.dxy.cn/ncovh5/view/pneumonia  Faster than worldometers for Chinese data, but slower on rest of the world data. I don’t think China cases matter anymore.

https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports I’m less enamoured of the WHO data now than I was at the start of the crisis. They are providing less information now than they were at the start of the crisis, and it sometimes contradicts country-level data.

https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6 The prettiest pictures, but one of the slower sites to update. I don’t find the charts that useful.

https://www.youtube.com/user/MEDCRAMvideos has a daily youtube wrap-up

https://www.youtube.com/user/ChrisMartensondotcom has a daily youtube wrap-up

Follow me


  1. QATAR Airways certainly doing their bit to spread the love… having said that most of the planes are flying empty so perhaps not too much love spreading

    • Qatar is just very good at testing – their mortality rate says that they don’t have much more cases per million than many many other countries

  2. Yes USA cases increasing.
    But USA deaths falling
    = lower case fatality rate over time

    Have we got some new improved treatment? No
    Has the virus mutated into a less dangerous form? No

    Is USA finding a greater % of total cases with increased testing capacity? Was USA previously diagnosing it’s more serious case with limited testing? Is USA incrementally finding more mildly unwell and asymptomatic cases with more widespread testing? Is this why the CFR (average risk of death) is progressively falling.

    Between middle April and middle May, USA cases fell from 30k to 23k. But USA deaths fell from 2k to 1k. And are continuing to fall despite “2nd wave.”

    • Yep. The IFR for this is about 0.6% based on 25 studies. May even be a bit less.

      The CDC symptomatic CFR is even lower at 0.4%.
      This includes all ages
      This meta-analysis included a few studies where the health system failed.

      If you are under 70, in good health, and are living with a functioning health system, you will be really really unlucky to get seriously unwell.

      COVID is not a nothing. And it is a straw-man argument of some here to suggest that a few of us are saying it is nothing.

      But it is not nearly as bad as it initially appeared.

      Right in cue, Beijing has reported another outbreak.

      • Steve, what about the excess deaths numbers we are seeing ? “…blood clotting, we see kidney damage, we see inflammation of the heart, we see stroke, we see encephalitis… this virus is probably a vasculotropic virus“.

        Also, separately, an early interview with a US ER Dr noted they had split the ER but were finding COVID on the non-COVID symptoms side. They were only getting tested incidentally. How many have died but likely are not attributed to COVID as not matching the early symptoms and never tested as such? Serious question.

        • All that you say about the pathophysiology is true. It has a lot of strange effects for a resp virus. What we thought we knew in February and March about viral pneumonia ….. we did not.

          1. The virus is pretty contagious. It spreads highly efficiently, especially in crowded environments – large apartment blocks, cruise ships, planes, jails.
          2. It is a new virus, meaning we are vulnerable. We initially believed we had zero immunity. There is some research suggesting a degree of cross-immunity from other coronaviruses, and so at a population level we may have a fair degree of immunity. This requires clarification.
          3. Due 1 and 2, it infected a lot of people very quickly. Maybe at a rate of 100,000 per day in USA. Maybe higher.
          4. New virus = new tests had to be developed and validated. The tests are not great, and miss many cases (maybe 1/3).
          5. There was a serious shortage of tests initially. There were strict criteria to get a test.
          6. We found the virus in the population we tested. We tested febrile returned travellers, it looked like a virus of febrile unwell travellers. But if we tested people with red hair or people with a last name starting with Z, then that is what we would have found.
          7. There is a very high % of people who get zero or minimal symptoms. Much higher than we understood. China knew this in February, but didn’t release that data about 21k asymptomatic people until March. For whatever reason, China kept the WHO and the rest of the world in the dark.
          8. These asymptomatic / minimally symptomatic people didn’t get tested. They didn’t qualify. These people would do very well, but they never got captured in the data.
          9. The sicker patients got tested. The more dire the situation in a country, the more this process distorted the true picture. So Italy was only testing the very sick – and the Italian data showed a high mortality.
          10. As testing capacity ramped up, the testing criteria were relaxed. And all these minimally unwell patients suddenly got caught up in the data. But this “silent majority” didn’t just appear when we started to look for them. They were always there.

          Notice how many VIPs got this virus. It was almost a VIP virus, and it was bad luck to be a VIP. Bit were VIPs more prone to getting infected? Or were testing criteria relaxed for VIPs? Who is going to tell Tom Hanks or a senior politician or a famous sportsperson they don’t meet the criteria?
          Yet how many VIPs died from this virus?
          That disconnect was an early suggestion the true picture wasn’t as it appeared.

          This virus does kill. A tiny %. But a tiny % of a huge population is a lot of death.
          Take NYC:
          8 million people
          25% have antibodies, so maybe 2 million infected.
          Say 2% get quite sick = 40,000 people
          Say 1% die (given system was overwhelmed and lots of chronically unwell in NYC) = 20,000 deaths. That’s a lot of extra deaths.

          To put this in context, approx 0.8% of NYC residents die per year. With Covid, 1% dead of 25% infected = 0.25%. That’s a big increment on a background annual rate of 0.8% – especially when compressed into a few months.

          Note there is not insignificant estimates with these figures. They represent my best sense of what probably went on.

          Your other point.
          We missed a lot of deaths. Maybe 50%. So double numerator.
          But we may have missed 80-90% of cases. So 5x or 10x denominator.
          That’s why the IFR is way down.

          • It is also worth having a read online about the experience of an increasing number of ‘recovered’ young COVID-19 patients. There a lot suffering from disabling symptoms – which appear to be a post viral inflammatory syndrome, perhaps similar to CFS or ME. There hasn’t been a lot of media reporting of this yet given the focus on deaths, but as the dust starts to settle I think we will be hearing about possibly millions of disabled younger people due to COVID-19.

          • Thanks for stepping me/us through that. It aligns to a few quite early ideas expressed here RE: the initial testing curves being a function more of the number of tests of an already present virus rather than a noiseless reflection of ‘r’ at that time. Some called for random samples for testing from the broader population at the time but these weren’t done (likely due to the shortage of tests I guess). The potential for ‘noise’ over the signal in the initial testing of an already present virus I though would have been called out by experienced epidemiologist with knowledge of big outbreaks. Anyway, it seems there is potential for big misses in both the numerator and the denominator for a new virus. If it is blood borne, I imagine they will be picking it apart for years to come – very novel. Anyway thanks for the considered response.

          • DeCODE in Iceland were one of the first groups to do random screening. They found a 1% prevalence in people randomly screened back in March. That seemed high for a fairly isolated island a long way from China. About 50% of these cases were asymptomatic.

            A small town (Vo) in Italy tested their entire population of 3300 in late February. 3% positive, but 50% of these were asymptomatic.

            That together with all the minimally afflicted VIPs suggested this was not a virus that killed 10%.

            We went through the same process with West Nile virus. Widespread concern with early suggestions of 10% mortality. Then the vastly greater number of asymptomatic cases were found. It turned out that < 1% were seriously affected, and even fewer died.

          • Ha, I was going to mention Iceland too but afraid it would be lost to the spam filter if the comment went too long. To be fair I in thought SARS 1 mortality was understated at the outset and ticked up? I thought this was the general trend from that article but will defer to you. The other thing about Iceland was some extension of their results which led them to assert that Italy may have kicked off via some very early spread via the UK but haven’t heard much on that angle since. I noted at the time because so many in the UK were looking sideways at transmission from the continent.

  3. So USA cases almost back up to where they were at the April peak. Yet deaths are down by 2/3.

    The truth is USA was experiencing WAY more than 30,000 cases per day. Way more. Which is how 25% of NYC got to have antibodies.

    They were only diagnosing a minuscule proportion. And it was not a random selection. There was a substantial spectrum bias such that they were diagnosing the sicker cases.