Vdci simulate the second wave. Save the smart quarantine of Gabriel’s door?

Will a smart quarantine, ie contact and route instructions, be enough to avert the kind of wave of coronavirus infection? The interdisciplinary tm vdc is trying to simulate it. Their city M has almost a thousand inhabitants. Hope that the result will be more accurate data.

He knows how to expand his knowledge. However, in the case of a new coronavirus pandemic, there is also a lack of data. The data itself is a lot, but reliable data errors. Researchers at the Model AntiCOVID-19 for R initiative presented to the public one of the attempts to overcome the current uncertainty. Based on available data, they create models that simulate how the Czech population is infected.

Are we modeling for? Because we have no other, explained Josef Lerka from the Faculty of Arts, Charles University, one of the speakers at the scientific symposium on May 19. We don’t have to experiment. We can’t close you Pardubice, leave everything open there and wait to see how many people get infected. According to him, a pandemic is an unusual event and our observational current events will make two sense and with a gap. According to him and according to his colleagues, the model is created.

On known data, the model can be painted to match the present. There are several models on which the results will correspond to our world. That doesn’t mean it’s a good model. But when you give a lot of different models, let’s say the press, they start to embrace some consolation that is worth paying attention to. And then on such models, scientists can carefully examine what the future holds.

do not knock on the first wave

Parameters of COVID-19 disease and v

Parameters of covid-19 disease and scientific uncertainty

One of the sources of data is logically the current literature. From these researchers drew data on how the new coronavirus SARS-CoV-2. Although it was published in the study, there is still no consensus on how basic the parameters are. According to Jan Trnka from the 3rd Faculty of Medicine, Charles University, it is therefore necessary to work statistically with this uncertainty.

But the virus is not the only factor in how it gets infected. Speed ​​en is the result of how people behave. Sociologist Daniel Prokop from PAQ Research examines a panel of 3,100 respondents every two weeks about a pandemic. The data reach into the ordinary, ie just before the first wave of measures. Using the diary method, respondents record what they do, how they work, how to meet km and how to read.


Development of social and risk activities

All types of activities were affected by the pandemic in some way, but of course some were not different. People have been dreaming of driving public transport and new restaurants (they were mostly closed). On the contrary, shopping in stores remained almost the same during the spring of 2020. The meeting with the family and the birds decreased and the end of the year, but during April it rose again and at the turn of May it was dark at the original level. I would not be in a hurry to believe that Easter was released at Easter, and Prokop points out to you that the people began to get married to others in the weeks before the Easter holidays.

Number of new confirmed orders in R daily (moving average)
date poet novch ppad
because nrst tdenn klouzav prmr
5.3. 4 2
6.3. 7 3
7.3. 7 4
8.3. 6 4
9.3. 8 5
10.3. 23 8
11.3. 31 12
12.3. 22 15
13.3. 34 19
14.3. 64 27
15.3. 84 38
16.3. 46 43
17.3. 90 53
18.3. 88 61
19.3. 172 83
20.3. 210 108
21.3. 91 112
22.3. 170 124
23.3. 71 127
24.3. 158 137
25.3. 260 162
26.3. 408 195
27.3. 217 196
28.3. 384 238
29.3. 166 238
30.3. 173 252
31.3. 306 273
1.4. 281 276
2.4. 269 257
3.4. 332 273
4.4. 282 258
5.4. 115 251
6.4. 235 260
7.4. 195 244
8.4. 295 246
9.4. 257 244
10.4. 163 220
11.4. 170 204
12.4. 89 201
13.4. 68 177
14.4. 82 161
15.4. 162 142
16.4. 130 123
17.4. 116 117
18.4. 105 107
19.4. 133 114
20.4. 127 122
21.4. 127 129
22.4. 95 119
23.4. 52 108
24.4. 85 103
25.4. 79 100
26.4. 52 88
27.4. 45 76
28.4. 55 66
29.4. 75 63
30.4. 103 71
1.5. 55 66
2.5. 18 58
3.5. 26 54
4.5. 38 53
5.5. 77 56
6.5. 78 56
7.5. 57 50
8.5. 46 49
9.5. 18 49
10.5. 28 49
11.5. 53 51
12.5. 45 46
13.5. 48 42
14.5. 82 46
15.5. 55 47
16.5. 49 51
17.5. 20 50
18.5. 111 59

Estimated number of meetings by type of work

Prokop considers the home office in limited risk contacts to be an interesting result. People who had the opportunity to work from home not only met me, but they also had other activities associated with the risk of infection. Going to work increases the number of people you lead regularly, and in eighteen weeks, Prokop remarked. Respondents thus responded to the extent to which coronavirus was exposed. Those who fought the least returned to work first.

Smart quarantine will work if…

Not all scientists have shown the new M model, recalling all the previous models A and B. Model A is based on the classical mathematical simulation of a command. People here are simply moved from the column of endangered to the field of infection and sweat according to the statistically described probabilities either among the asymptomatic, sick, recovered, hospitalized or died.

Model SEIR

SEIR model individual states


The course of the disease

The model is tuned to the well-known information about the course of the disease in R. According to economist Eva Hromdkov (NB, CERGE-EI), the model A has two and surprisingly accurate results. This, of course, is not a sign that future predictions will work.


Short-term pedpov model A

Before the month, we said almost the same thing: the flatbed works, but if we don’t replace them, they will have to be tightened again, Hromdkov remarked.

In the medium-term insight, the model’s cars showed you three scenes, depending on how the route and contact will work (the so-called smart quarantine), or how the measures will be received.

This is what the situation would look like if the plan in the Czech Republic remained for several more months:


Pedpov pokraovn plonch opaten

Assuming that the government is about to release, this scene will not occur either. Depending on the model, the loosening itself would result in a gradual increase in exponential series (note the changes in the scale of the vertical axis). We would get to 250 cases a day sometime at the end of August. In June, it would be every day a thousand thousand new. Of course, no one would let it go that far, and a measure would be introduced again with the aim of reducing orders.


Pedpov without measures, at the same time level route

but would it be even without the kind of wave limited? Hromdkov, based on the model, thinks so. However, it is necessary to increase the activity of routing and supervisory contact with hygienic stations (I can help with the implementation of smart quarantine). Even a slight change in the route performance parameter completely reverses the course of the command:


Prepaid without measures, increase the routing

First, there are good models. Two ideas about how big or in this case the change should have affected the overall situation. But the A model is indisputably limited. The model does not take into account the breeding people, Hormdkov pointed out. This is the attempt and the M model, which was presented by researchers from the antiCOVID-19 initiative.

Msto pln aktivnch lid s pbhem

Let’s say you have to simulate a city for a round. How would you do that? Would you try to build virtual houses, place virtual people in them and program some breeding for them? Interdisciplinary darkness a little more important. People are not robots, it is important for them to meet, family, friends, leisure, procurement, etc. Instead of the rest of the graphical simulator, they used the theory of social networks (Social Network Analysis, map structures of web, confusion with web services).


Analza socilnch st v modelu M

Each hunter is a knot in this model, explained Tom Divik from the Faculty of Arts, Charles University. But just as real people don’t have just one type of relationship, so imaginary people weren’t a node in just one graph. Instead, each hunter had a bite and a few charts. These graphs represented his involvement in society: in the family, in the cola, in his friends, colleagues, shoppers, etc.

Parameters that
Parameters that
Parameters that
Parameters that

Parameters that affect the speed of en

In different layers, the infection changes differently quickly. It is the bottom of you because, in different environments, people are far and long away from each other. The researchers estimated the following probabilities based on the following addressed expert:

Where is the best

Where the greatest simulation is simulated

According to them, the most important is long-term contact for a short distance, you stay in one household. On the contrary, moles in the open air are estimated to pose the least risk.

Where is the best

Where the greatest simulation is simulated

In total, there are about 56 thousand people in the virtual city of M, Martin showed from the state of information theory and automation of the Academy of Sciences. so ask for the composition of what happened during the pandemic.

Model M

City parameters M

On this model, it is then possible to simulate (and simulate against relative data) even complex measures, such as wheel suspension, no rouek or various functionality or malfunction of the route (so-called smart quarantine).

Gabriela-29691 z m

Gabriela-29691 from the city of M

The model goes and on the level of individual people. The researchers showed you nineteen-year-old Gabriel-29691 ″. At the request of the newspaper, they were able to specify how long this virtual house lives (shared by a household with two women aged 50+) and how often it is known (it is part of a family of six). In addition, the model has the ability to generate pbhy, we can also watch how Gabriela in one of the scenes went to the verek where the covid-19 became infected. She learned a few days later that her friend, who had also become infected, reported her as one of his contacts in creating the memories of the map.

Gabriela-29691 is going to a party

Gabriela-29691 is going to a party

Such detailed pbhy are just interesting, show the details to which the model goes. But we are interested to know that the M model will happen when it comes to the future. The results are generally similar to those in Model A: quarantine, if it works for a long time, can reduce the reproduction below 1, thus preventing the number of epidemics.


The course of infections with the SARS-CoV-2 virus in the Czech Republic according to the M. model. Oranov is a real possibility without any measures, but with a marked change of breeding. Green is the result with limited walls, blue is the development with the preservation of all areas.

Model M generates a scn printer with various parameters. Even a small difference over time causes large deviations, so the graphs can work with certain uncertainties and probabilities (this is the degree of uncertainty shown on the graph by scattering around the plotted curve).

According to medical expert and economist Pavel Hrobon, who closed the two-hour written Friday, the Czech Republic managed the first wave of the epidemic. According to him, models can help to decide on the basis of specific values ​​in the future: It is a better decision-making in case of a kind of epidemic wave. In the first wave, we had to act fast. But we can act a lot.

We handled the first wave perfectly, but the price was high. Klov is not allowing the second wave, Hrobo emphasized. But the probability of a kind of wave according to expert estimates is not negligible.

We need two basic things for that. The tools are therefore smart quarantine and information. According to him, the government should first decide on further measures to store data and information. Models can provide much more accurate information.

According to Hrobon, if the second wave comes and new measures are needed, it should be much more washed, much more selective: If, as a company, we cannot prepare for the second wave, it would be completely unnecessary and, in my opinion, unforgivable.