Projections update

 
The effectiveness of our module which projects the evolution of COVID-19 in CH is defined by the quality of its projections which in turn depends on the quality of the data it receives. Please note that these forecasts (as does any forecast) depend on the quality and magnitude of the data. What we show reflects the best we can do with what data we can collect in the current situation.
 
In this context, we found that when comparing what the BAG (Swiss Federal Ministry of Health) declares with data available on a Swiss canton level, consistency is not guaranteed: the BAG works with a ratio of cases per 100’000 inhabitants, while other data are often heterogeneous extrapolations.
 
This is why we decided to create a more coherent data set and to rely on it to redo our COVID-19 in CH.

COVID-19 in CH

As a reminder, we display 3 key indicators n the interactive maps down below:
 

1. The Blue Circles are Healthcare facilities
Their sizes scale with the number of beds reported in 2017 here

 

2. The Red Circles illustrate the total reported Covid-19 cases per Canton centered on the corresponding “capital” (main city)
Their sizes, like in the John Hopkins map are relative to the number of cases

 

3. The lines are the central point of the module : the more visible they are, the more relevant the (hospital) choice.
Our module finds the most relevant dispatching hospital (destination) for a given COVID-19 infection cluster depending on proximity, number of beds and number of new cases

 
We therefore “reconstructed” the population as presented by the BAG and on this basis we relaunched our model. Here is the update of the projections COVID-19 in CH with data compatible with those of the BAG:
 

Switzerland COVID-19 situation on 29.03.2020

 
What stands out: with this “revised” data set, the evolutionary trend seems to show that the epidemic is stronger in the French-speaking part of Switzerland than in the German region (is it the result of voluntary containment measures which are beginning to produce results or is it linked to the density of health structures? Probably a bit of both). Some French-speaking cantons seem less affected than others, but they do not have the same population density either. Finally, our projections seem to show that the progression of COVID-19 in VAUD remains significant.
 
Our module shows the forecasts for the next few days . The map below illustrates the situation to be expected in 3 days’ time (i.e. on 02.04.2020) and the complexity that it could signify in the most affected areas.
 

Switzerland COVID-19 situation on 02.04.2020 (forecast)

COVID-19 in CH : focus on Zürich

The canton of Zurich is one of the most interesting area given its demography and its infrastructure :
 

Zurich COVID-19 situation on 29.03.2020

 
Based on the available data, the situation could change as follows :
 
COVID-19 in CH : focus sur Zürich : 29.03.2020 (by CALYPS)
 
What stands out: the blue dots are the current data (according to our information). The different models provide an idea of ​​the possible evolutions: in orange the exponential model; in red the polynomial model; in green the SIR model and in purple the model closest to the data. We note that Zurich appears to be in an area between exponential and SIR. This means that we could likely be approaching the peak of the epidemic in this region.
 
As you now know if you read our blogs on COVID-19, our module displays forecasts for the next few days and the map below illustrates the situation that can be expected on 02.04.2020 in the canton of Zürich :
 

Zürich COVID-19 situation on 02.04.2020 (forecast)

COVID-19 in CH : focus on Vaud

Let’s now focus on the canton of Vaud, which is also one of the most interesting areas given its demographics and its infrastructure :
 

Vaud COVID-19 situation on 29.03.2020

 
Based on the available data, the situation could change as follows :
 
COVID-19 in CH : focus sur Vaud : 29.03.2020 (by CALYPS)
 
What stands out: the blue dots are the actual data (according to our information). The different models give an idea of the possible evolutions: in orange the exponential model (which is flat given the data variability); in red the polynomial model; in green the SIR model and in purple the model closest to the data. It can be seen that Vaud is in a very uncertain area between the polynomial and SIR models. This indicates that the peak is most likely not yet reached and that the virus should still continue to spread. Note that the data reported for the canton of Vaud have a very high variability, which leads to a bias in the models.
 
Here are the forecasts for the next few days. The map below illustrates the situation that could be expected on 02.04.2020 in the canton of Vaud :
 

Vaud COVID-19 situation on 02.04.2020 (forecast)

 
Of course, we welcome any comments on the accuracy of the map data we represent. Indeed, since they come from several data sets, small errors can never be excluded. So do not hesitate to write to us if something is not right.

Anticipating the impact of COVID-19

CALYPS can help with the looming coordination challenge, mainly by helping to better anticipate impending flows, provided data is available. At present this is a major problem in Switzerland, due to data incompleteness, non-alignment and non-standardization, whether in and among municipalities, cantons or linguistic regions.

 

Our artificial intelligence platform is specialized in flow prediction thereby enabling superior anticipation. The results it generates reflect the harsh reality of the field and its good performance has been demonstrably verified. By feeding it with relevant data (as described above) and by using its predictive capabilities, we are convinced that it could help authorities and healthcare institutions to better understand the impacts of the COVID-19 pandemic.

 

To all of you who read this BLOG, do not hesitate to share it and relay it to all the people likely to use it: in this fight, all help is precious, time plays a key role and good anticipation means being that crucial step ahead !

 
Team CALYPS
 

Responsibility waiver regarding CALYPS predictive information : the accuracy of any forecasts or predictive information depends critically on the quality, timing and sample size of the data. The same applies to CALYPS generated forecasts which are intended as an aid in helping to anticipate . CALYPS does not assume any responsibility or liability for any conclusions and decisions reached, including possible ensuing plans taken by any party in its use of CALYPS information. 

Your data is valuable!

Data Intelligence is intended for managers, decision-makers and business leaders. It is the means, the tools and the methods designed to provide valuable information to take a decision and act. Put all the chances on your side by relying on data professionals.

CALYPS experts are here to help you