Coronavirus COVID-19 outbreak statistics and forecast v0.86



Most plots are updated daily. Downloading the latest data ......

This chart shows the total number of cases and deaths in different countries. Italy and Iran quickly surpassed South Korea as the country with the most cases outside China. South Korea does an excellent job of rapidly testing and slowing the growth significantly. Many heavily infected European countries have quickly growing COVID-19 cases.


This interactive plot is better viewed in the log-scale (check the button in the China tab). Scroll to the right to see all countries. Single-click the name of a country to select or deselect the corresponding curve. Double click a name hides all others. Many states now in the rapid, exponential growth phase, especially Italy, Iran, France, Spain, Germany, and the U.S.


The death tolls are increasing sharply in many countries, especially in Italy and Iran. Italy may have more deaths than China in a few weeks. This is better viewed on the log-scale.


This interactive plot shows the trajectories of growth in different countries, plotted the number of confirmed cases by the number of days since the 100th confirmed case. Select an area to zoom in and mouse over to see the name of the country. Except for Singapore, Bahrain, and Japan, which have slower growth rates, the rest of the states are on a similar exponential growth trajectory. Iran and Italy are showing signs of slowing down in growth rates. But Spain and the U.S. are still growthing rapidly. Continuing this trend, both countries could have millions of cases in weeks. Unlike China, the U.S. is seeded broadly early on. Many clusters, like Washington state, New York, California, already formed. Each could balloon into a Wuhan-sized outbreak.


This is a static version of the plot above. The virus is now spreading faster in Europe and the U.S. than in Asia. In many Asian countries, especially Japan, people routinely wear face masks when they are sick, so that they do not infect others.

The top 30 most affected countries are plotted by total confirmed cases (x-axis) by the current daily growth rate (y-axis). The growth rates are estimated by a regression analysis (linear model) using data from the last 7 days. The growth rates are the slopes observed in the figure above for each country. In the top right are the most serious countries (Spain, the U.S., Germany, France, U.K., Switzerland) with more cases and still rapidly growing. China has the most confirmed cases but is barely growing. Portugal has the highest growth rate. Countries in red have higher death rates.


The death rates in Germany, South Korea, Switzerland are lower than these observed in China and Iran. The death rate in Italy is highest, probably due to under-diagnosis.


Data from the New York Times.
Please wait while we retrieve today's data. All analyses on this page are based on data from this R package by Rami Krispin.
This shows the confirmed cases in the top 20 states/provinces with the most confirmed cases.


This map shows that the clusters of COVID-19 outbreaks on the east and west coast. Washington State, New York, and California lead the nation.


The daily growth rate is calculated based on the last seven days. States in dark grey do not have enough data points. If the growth rate is 26%, expect confirmed cases to double every three days. If the growth rate is 42%, expect confirmed cases to double every other day. Many states fall between 20% and 40%. The alarmingly rapid spread of COVID-19 is observed in Michigan, Louisiana, Ohio, Utah, Connecticut, Minnesota, Wisconsin, Florida, and New Jersey.


This interactive plot shows how the trend in each of the top states. We recommend viewing this in the log-scale by going back to the China tab and click on the radio button. The exponential growth is apparent as we see a straight line in the log scale. Massachusetts and Oregon seem to have a slower rate of increase over other states that share a similar slope. Single-click a state name in the legend to show or hide the corresponding curve. Double-click a state to hide all others. But COVID-19 infection is across the U.S. Thanks to exponential growth, even states with just a single patient could explode to hundreds of thousands of cases in a matter of weeks.


The provinces/states with 20 or more cases are lined up by the day that they reached 20 cases. In the log-scale, we can see what provinces are having a faster increase from the slope of the curves.


The most affected provinces are plotted by total confirmed cases (x-axis) by the current daily growth rate (y-axis). The growth rates are estimated by a regression analysis (linear model) using data from the last seven days. The growth rates are the slopes observed in the figure above for each province. In the top right are the most severe countries with more cases and still rapidly growing. If the growth rate is 26%, expect confirmed cases to double every three days. If the growth rate is 42%, expect confirmed cases to double every other day.


We used a simple time-series forecasting model provided by the forecast package in R and the exponential smoothing method. The parameters are ets(data, model="AAN", damped=FALSE). Only the last 21 days of data are used. Data are first log-transformed. The linear projection is based on the assumption that the log-linear trend holds in the near future. It is more accurate for the near term, i.e. the next few days. Einstein reportedly said that compound interest is the most powerful force in the universe, and those who do not understand it pay it. If we have 26% more cases than yesterday, then it doubles every three days. This means a 100-fold increase in cases in 20 days! When preparing for a pandemic, public officials need to grasp the power of exponential growth. They should pass a math test before assuming their offices.


We used a simple time-series forecasting model provided by the forecast package in R and the exponential smoothing method. The parameters are ets(data, model="AAN", damped=FALSE). Only the last 21 days of data are used. Data are first log-transformed. The linear projection is based on the assumption that the log-linear trend holds in the near future. It is more accurate for the near term, i.e. the next few days. Einstein reportedly said that compound interest is the most powerful force in the universe, and those who do not understand it pay it. If we have 26% more cases than yesterday, then it doubles every three days. This means a 100-fold increase in cases in 20 days! When preparing for a pandemic, public officials need to grasp the power of exponential growth. They should pass a math test before assuming the office.


To make the data stationary, we transformed the data into daily percentage increases of confirmed cases by country, and run the forecast. The parameters are ets(data, model="AAN"). For some countries, the daily percentage changes fluctuate widely, making the modeling difficult.


The cumulative deaths due to COVID-19 is forecasted. The parameters are ets(data, model="AAN", damped=FALSE). Only the last 21 days of data are used. Data are first log-transformed. Total deaths are also growing exponentially.


Loading county-level historical data from the New York Times.
The counties with 20 or more cases are lined up by the day that they reached 20 cases. We can see which county is having a faster increase from the slope of the curves. Note that this is on the log-scale. Most counties should expect cases to double every 3 days. That means in 30 days there will be a 1024-fold increase in health care needs.


This barplot shows the number of confirmed cases in the counties in the selected state. The number of deaths is in the parentheses.


This map shows the current number of cases by county.


This map shows the daily growth rates by county. The growth rates are estimated by a regression analysis (log-linear model) using data from the last 7 days.


The counties are plotted by total confirmed cases (x-axis) by the daily growth rate (y-axis). The growth rates are estimated by a regression analysis (log-linear model) using data from the last 7 days. The growth rates are the slopes observed in the figure above for each county. In the top right are the counties with more cases and still rapidly growing. Highlighted in red are counties with higher death rates.


Projections for each county. Note that this is a very rough estimate based on the log-linear model. We should be especially cautious when we extrapolate over a week, as the trend could change.


All analyses on this page are based on code and data from this R package by Rami Krispin.






China total confirmed: 85485 (+40, 0%), suspected:8 (0, 0%), death:4648 (0, 0%), discharaged:80279 (+11, 0%)

Updated July 10 Beijing time 17:19:30 . New cases may not be final count for the day. Not optimized for mobile phones.
In just ten days from Jan. 15-25, 2020, the coronavirus spread from Wuhan to all over China, a vast country connected by high-speed trains. It was the perfect timing for the virus. Jan. 25 was the Chinese New Year. About 3 billion trips would be made, the largest human migration on the planet.


Most of the cases are in Hubei Province. On Jan. 23, China took an unprecedented, decisive action to put about 1 billion people under quarantine. Although controversial, this effectively contains the virus to Hubei. It could be much (100x) worse, as each of the densely populated metropolitan areas like Beijing and Shanghai could explode. They still could, as people are going back to work.

Three weeks after the lockdown, total cases stabilized around Feb. 20. On Feb. 22, the confirmed cases jumped by 15,139 as the diagnosis guideline changed from relying solely on the nucleic acid test to include all suspected cases, namely patients with clinical features (lower lung infection on CT scan). This is an interactive plot, where you can mouse over or zoom in. Also, you can change the linear coordinate to log-scale, which makes more sense as the cases grow exponentially.

This plots the percentage of daily increases from the previous day. It compounds by 30% to 40% daily in the early stage. The rapid, exponential growth phase in China spans roughly from Jan. 15 to Feb. 15, 2020, when the number of confirmed cases skyrocketed 1670-fold from 41 to 68,500. Such rapid growth is now evident in many countries: Italy, Iran, France, Spain, Germany, France, and the U.S. Public health officials need to grasp the power of exponential growth.

This chart shows how the total number of cases increased in the provinces outside Hubei. Since Feb. 15, most areas do not see an increase or decrease. With a total of 879 confirmed cases, China implemented a massive lockdown on Jan. 23. But cases continue to soar 78-fold in the next three weeks. There are now at least ten countries across the world with a critical mass of 879 or more cases. None acted as decisively. All these countries must prepare for a 100-fold increase in COVID-19 cases.

Each point represents a Chinese city with 200 cases or more. Colored by provinces. Mouse over to see the name of the cities. Select an area to zoom in the lower left. Cities above the trend line have higher mortality. Probably due to an overwhelmed healthcare system, the death rate in Wuhan (excluded from this plot) is much higher than the average in other cities.

The gross death rates are estimated by dividing the current total deaths by total confirmed cases. Cities in Hubei province have higher fatality rates than cities in other regions.




July 10 Downloading map......

Scroll down to see world map.

For feedbacks or suggestions please contact me via email or Twitter. My research interests are genomics, bioinformatics, and data science (lab homepage). Source code on GitHub. I am not a epidemiologists or statistician, so be critical of my analyses. I am just a college professor having fun in a basement during spring break. But I am enjoying it more than my students on the Florida beach ...
Accuracy not guaranteed. Part of the data is not official statistics.
For details on data sources see our preprint.
This website visualizes the data and trend of the 2019-nCoV (SARS-Cov-2) coronavirus pandemic. Developed based on these R packages: nCov2019 by Dr. Guangchuang Yu, coronavirus and covid19Italy by Rami Krispin. U.S. state and county level data from the New York Times.

不保证数据和分析的可靠性,仅供参考。

该网站是我工作之余仓促码出来的, 难免有错误。见 源代码。 主要目的是帮助朋友们了解疫情。纯粹个人行为。 bcloud.org 是以前注册的一个域名,随手拿来用了,不属于任何组织。
之所以能很快写出来,最主要是因为南方医科大学的 余光创教授 (微信公众号biobabble)写了一个功能强大的下载实时数据的软件包: nCov2019。 实时数据来自腾讯, 每天更新。历史数据来源:丁香园,国家卫健委,和 GitHub. 数据每天更新。
有意见或建议可以给我发 邮件 。 我做生物信息学方面的研究,用计算的方法探索生命的奥秘 ( 研究室网页 )。
武汉加油! 中国加油!
2/5/20 Version 0
2/8/20 Version 0.1
2/9/20 Version 0.2
2/12/20 V 0.3 English version
2/23/20 v. 0.4 Interactive plots.
3/12/20 v. 0.5 Historical trend among countries.
3/15/20 V. 0.6 Changed forecasting parameters from default to ANN.
3/16/20 V. 0.7 Add provincial level data for U.S. and other countries based on the coronavirus package.
3/20/20 V. 0.8 Add Italy data
3/27/20 V. 0.8 Add us data from the New York Times.
3/28/20 V. 0.8 Add detailed county level data from the New York Times.