COVID-19 Forecast for the Post-Soviet Space

/in , , , , , , , , /

Data Source:
https://github.com/datasets/covid-19

The forecast methodology is developed by the author.

Any real-life exponential growth can be described by a logistic curve. The developed method is based on the time series. Based on the growth trajectory, it calculates the logistic growth rate or the steepness of the curve, which in other equal conditions does not change through the course of the growth and continues the growth path. The model assumes that the current quarantine measures will be unchanged till the end of the pandemic. This model can’t predict the next waives. Thus, if the social distancing measures are loosened there is a possibility of the next wave.
The developed method uses the “days to double” measure which can be seen in the lower right chart of each country to calculate the projection. This measure shows how many days will the total cases double. The exponential growth has both a surprisingly sudden acceleration and surprisingly fast downfall, as it can be seen on the example of China, South Korea, Australia, and some European countries.
The coefficient of determination or R2, which can be seen on each lower right chart, shows how much the model explains the reality. In this case, the coefficient of determination can explain the noise in the data. If the testing is not consistent, then R2 will show it. The other reason for the low R2 is a prominent change in social behavior.

The forecast is validated on the countries that have passed the inflection point. It should be noted that even if the data is incorrect and the measuring scale is broken, the more important is the progression or the change. In general, if the coefficient of determination is higher than the progression of the test results mirrors the reality, in general, regardless of which percentage of the real cases is tested.

This progression can show when the peak or the inflection point is expected and can assist the policymakers to decide when it is safe to loosen quarantine measures. It must be stressed out that it is only safe to open the economic activities if there are no new inside cases in the country or the territory, otherwise there is a chance of the second wave, which can be more devastating than the first one. It can be safe to say that the schools will remain shut until there are no new inside cases. There is a high probability that international travel may be limited until the end of summer. This can cause further constrictions to the financial liquidity of some countries.

The countries are presented in alphabetical order.

COVID-19 in Armenia

Armenia is experiencing the second wave of the pandemic. The second wave started around April 12th. The inflection point or the peak of this wave is expected at the beginning of May. It must be noted that the growth in cases is registered mainly outside of Yerevan. The cases in Yerevan are on the decline and some quarantine measures in Yerevan are already loosened. The coefficient of determination in Armenia was quite high before the beginning of the second wave, which speaks about the quality of the data. Presently this number is low because the dynamics are bimodal. The forecast is calculated based on the dynamics of the second wave.

 

COVID-19 in Azerbaijan

Azerbaijan has passed the peak and presently is on the decline. The government of Azerbaijan should continue the strict measures as there is a high risk of the second wave if the measures are loosened. These measures should be continued until the last week of May, given there is no second wave.

 

COVID-19 in Belarus

Belarus has one of the most peculiar dynamics. Till April 12th it took around 3 days for the number of total cases to double. On April 12th there were 2,578 cases in Belarus. If this progression was continued then on April 15th there should have been around 5,000 cases and around 10,000 then 20,000 cases on April 18th and 21st respectively. The continuation of the progression, which is based on the given statistics, shows that the peak has been already reached or will be reached soon. This sudden alteration is explained if there is a change in social behavior. The Yandex self-isolation index does not show a noticeable change in social behavior in Minsk[1]. The next two weeks will show the real dynamics in Belarus.

COVID-19 in Georgia

The growth speed of Georgia is an outlier since it has a steady doubling rate of around 10 days, which hasn’t changed much in recent days. The issue with the progression is that it has a slow steady growth, thus there will not be a lot of registered cases, but the curve is comparatively more spread out. Thus, the economy might be locked up for more than 3 months. Which may increase the negative effect on the economy of Georgia. Georgia has one of the strictest lockdowns of the post-soviet countries, Аlthough the number of daily cases is not that high, the strict measures should result in the decline of cases.

COVID-19 in Kazakhstan

A mini-wave of the infection can be noticed in the data of Kazakhstan. This mini-wave has started on April 12th with the local peak registered around April 17th. It should be noted that there is a lot of uncertainty in the data which is shown by the low coefficient of determination. The forecast for the separate regions of Kazakhstan should be carried out to reduce the uncertainty and make the forecast more reliable.

COVID-19 in Kyrgyzstan

Kyrgyzstan is on a good trajectory. It has passed the inflection point. If there are no new waves and the present social practices will remain till the end of May than Kyrgyzstan is expected to have around 800 cases in total.

COVID-19 in Moldova

Moldova shows a sharp change in social behavior around April 6th, which reversed the growth pattern. If there are no new waves and the present social practices will remain till the end of May than Moldova is expected to have around 4,000 total cases.

COVID-19 in Russia

Presently it is pervasive to give a forecast based on total cases in Russia. The country is very large. Each subject of the Russian Federation has its own dynamics and there is a lot of uncertainty in the cumulative cases. Because of that, only the forecast for Moscow can be given. Of all the forecasts of the Russian subjects, Moscow has the largest coefficient of determination, since the first case in Moscow was recorded earlier than in other subjects and Moscow has the longest time series. The dynamics of the pandemic in Moscow resemble those of New York. At the peak, about 6,000 registered patients are expected daily, which according to the current forecast will be the case around the middle of May.

40% of all identified cases with COVID-19 in Russia are registered in Moscow. This ratio is not prawned to change throughout the pandemic. Using the projection of this percentage the total and daily numbers for the Russian Federation can be calculated. In general, the other subjects of Russia presently have a sharp growth rate. The quarantine measures should be tightened even if there are a few registered cases. The sooner the social distancing measures are implemented the sooner the economy may open up again.

COVID-19 in Ukraine

The projection shows that the peak will be reached soon. If the current measures will persist till the end of June and no other waves are recorded than there will be around 20,000 registered cases at the end of this wave. Ukraine will be able to fully open its economic activity at mid-July.

COVID-19 in Uzbekistan.

Uzbekistan is on a good trajectory. It has passed the inflection point. If there are no new waves and the present social practices will remain till the end of May than Uzbekistan is expected to have around 3,000 total cases.

Among post-soviet countries, there are no recorded cases in Tajikistan and Turkmenistan. Thus, it is hard to present the real situation there and give a forecast.

 

[1] https://yandex.ru/maps/covid19/isolation