Predict Epidemic Dynamics of COVID-19

English / 日本語
(How to read the graph/ 予測結果の見方)
(Update information)

We found a technical trouble causing the graph below to be not visible (June 25th). We have fixed it (June 25th).

<HOW TO read this figure> (Visual instruction and explanation on prediction is here)

  • This figure shows the cumulative number or the 7-day averaged daily number of infected cases from January 22nd to the latest data, and also shows its prediction on the next 100 days from the latest day analyzed. (We will update almost everyday)
  • We analyzed the countries or regions reported over 10,000 cumulative infected cases on the latest reported day. The prediction for each country includes the median estimate and the 95% credible interval (shown, if mouse-over each curve).
  • The reported data and its prediction are shown in lighter and darker points respectively (at the left- and right-hand side of the gray bar).
  • You can choose a particular set of countries/ region/ N-largest or N-smallest number in data or prediction, out of the country selection box underneath of the figure.
  • You can also adjust the dates to show the plot.
  • You can choose one of the two ways of visualize the graph, in the linear or semi-logarithm plot, by pressing the radio button at the top of the country selection box.
  • English and Japanese version are available. (日本語版はこちら)

<About Use or Citation of This Website>

  • This is by no means the “truth” but merely a prediction estimated from a part of data of COVID-19.
  • The technical method to make this prediction has NOT been approved, but will be or is submitted to an academic journal.
  • We do NOT take any responsibility to any of your personal or industrial usage of this website. Please use this result at your own risk.
  • The method or the predicted values may be changed or updated without any prior notification.
  • Please cite the link or preprint below, if you use or cite this prediction in this website.
    Citation (web): http://www.jaist.ac.jp/project/prepidemics/
    Citation (preprint): Hidaka, S. & Torii, T. (2020). Predicting Long-term Evolution of COVID-19 by On-going Data using Bayesian Susceptible-Infected-Removed Model. MedRxiv 2020.05.08.20094953. 
    Project Prepidemics developed by Shohei Hidaka, Associate Professor in Japan Advanced Institute of Science and Technology (JAIST).
    Graphic design by Takuma Torii, Assistant Professor in Japan Advanced Institute of Science and Technology (JAIST).
  • Contact us by email: prepidemics(at)gmail.com.

<Outline of the method of prediction>

  • We analyzed the timeseries of the cumulative number of infected cases for each country from the earliest day with more than 100 cases since January 22nd, 2020, to the latest day of report. We applied a novel technique based on a Bayesian extension of the Susceptible-Infected-Removed model to the data (provided by Johns Hopkins University CSSE cited below), and estimate the parameters in the epidemic dynamics, which is used to generate the predicted timeseries illustrated in the figure. Please read our preprint here for more detail, which is a pre-published academic paper, to learn more detail about the method.

This prediction was made on basis of the data repository provided by Johns Hopkins CSSE <https://github.com/CSSEGISandData/COVID-19>.