Skip to content

2020ΒΆ

  • 1 min read

Comparing Rt numbers for Belgium on 09-12-2020

Model Based on URL Rt Date
by Niel Hens Cases https://gjbex.github.io/DSI_UHasselt_covid_dashboard/ 0.96 20-12-5
Cori et al. (2013) Hospitalisations https://covid-19.sciensano.be/sites/default/files/Covid19/COVID-19_Weekly_report_NL.pdf 0.798 20-11-27/ till 20-12-3
RKI Hospitalisations https://datastudio.google.com/embed/u/0/reporting/c14a5cfc-cab7-4812-848c-0369173148ab/page/ZwmOB 0.97 20-12-09
rtlive Cases https://rtlive.de/global.html 0.80 20-12-09
epiforecast Cases and Deaths https://epiforecasts.io/covid/posts/national/belgium/ 0.5 20-12-07
Huisman et al. (2020) Cases https://ibz-shiny.ethz.ch/covid-19-re-international/ 1.01 20-11-24
Huisman et al. (2020) Hospitalisations https://ibz-shiny.ethz.ch/covid-19-re-international/ 0.84 20-11-24
RKI Cases https://twitter.com/BartMesuere/status/1336565641764089856 0.99 20-12-08
Deforche (2020) Hospitalisations and Deaths https://twitter.com/houterkabouter/status/1336582281994055680 0.85 20-12-09
SEIR Hospitalisations and Deaths https://twitter.com/vdwnico/status/1336557572254552065 1.5 20-12-09
  • 8 min read

Estimating the effective reproduction number in Belgium with the RKI method

Using the Robert Koch Institute method with serial interval of 4.

Every day Bart Mesuere tweets a nice dashboard with current numbers about Covid-19 in Belgium. This was the tweet on Wednesday 20/11/04:

twitter: https://twitter.com/BartMesuere/status/1323881489864548352

It's nice to see that the effective reproduction number (\(Re(t)\)) is again below one. That means the power of virus is declining and the number of infection will start to lower. This occured first on Tuesday 2020/11/3:

twitter: https://twitter.com/BartMesuere/status/1323519613855059968

I estimated the \(Re(t)\) earlier with rt.live model in this notebook. There the \(Re(t)\) was still estimated to be above one. Michael Osthege replied with a simulation results with furter improved model:

twitter: https://twitter.com/theCake/status/1323211910481874944

In that estimation, the \(Re(t)\) was also not yet heading below one at the end of october.

In this notebook, we will implement a calculation based on the method of the Robert Koch Institute. The method is described and programmed in R in this blog post.

In that blogpost there's a link to a website with estimations for most places in the world The estimation for Belgium is here

LSHTM

According to that calculation, \(Re(t)\) is already below zero for some days.

  • 3 min read

My talk at data science leuven

Links to video and slides of the talk and thanking people.

Talk material

On 23 April 2020, I was invited for a talk at Data science Leuven. I talked about how you can explore and explain the results of a clustering exercise. The target audience are data scientists that that have notions of how to cluster data and that want to improve their skills.

  • 4 min read

Evolution of burglary in Leuven. Is the trend downwards ?

Evolution of burglary in Leuven. Is the trend downwards ?

The local police shared a graph with the number of break-ins in Leuven per year. The article shows a graph with a downwards trendline. Can we conclude that the number of breakins is showing a downward trend based on those numbers? Let's construct a dataframe with the data from the graph.