Covid Example

Analyzing Worldwide Immunity and Vaccination Progress of COVID-19

Analyzing Worldwide Immunity and Vaccination Progress of COVID-19

The overall purpose of the data analysis is to measure progress towards COVID-19 pandemic ranging from the start of vaccination until the end of April 2021. For that regard, we have selected different datasets.

The first one gives us an overview of vaccinations by day and country and type of vaccines administered. The second one offers a more detailed progress with herd immunity over type and people fully vaccinated per hundred , daily vaccines administrated.

Last but not least we crossed this joined two datasets with the Economic dataset of countries to analyze economic factors that influence vaccination progress. We have tracked several parameters focusing overall in people fully vaccinated per hundred as a signal of immunity, being defined as the ratio(%) between population immunized and total population up to the date in the country.

01. Herd Immunity progress

Find in the right side map the insight of people fully vaccinated per hundred per country in more than 150 countries, crossed with a key metric (GPD Growth) from countries with respect to herd immunity progress. 


The visualization aims to respond to the question: Which are the countries that in general are more advanced with respect immunity? 


From this visualization and overall general max values of the dataset we can extract the conclusion that the Leaders in herd immunity are: Israel, United States and Chile.

02. Evolution of herd immunity over time

In the left side graph we plot the herd immunity over time comparing the top ten countries that we extract from the map above. In this case, note that we have selected people fully vaccinated per hundred and not daily vaccinations administered because some of the vaccines are bounded to several dosis. Find plotted below the top ten countries and their growth.

Seyshelles & Israel have developed almost 60 people fully vaccinated per hundred, almost doubling Chile, Bahrain, United States, San Marino, that are in between 35 and 25 people fully vaccinated per hundred. Malta, Wales, Scotland and Northen Ireland are around 25 people fully vaccinated per hundred.

Having into account that Sheyshelles is a small archipielagus of 90K inhabitants, the vaccine scalation strategy might not be as hard to implement as another countries such as Israel and Chile, with 9M and 19M respectivelly.

Note that Israel and United States show a logistic growth, being key for the autoregressive model growth prediction. Other leader countries follow these logistic patterns (Chile), although linear growth is also in between this case. (Germany, France, Italy), setting a notable difference regarding immunity progress among them.

03. Factors that influence vaccination

Which economic factors influence immunity progress rhythm?

This section aims to give a hint to the question which economic factors influence immunity progress rhythm? 

 We use Pearson correlation coefficient, which is a measure of the linear association between 2 variables. It ranges from 1 to -1 giving us the idea of perfectly negative linear correlation in between two variables, no linear correlation between two variables and a perfectly positive linear correlation between two variables

The correlation matrix including data from 41 countries is shown into the right graph.

The column we are interested in is people_fully_vaccinated and people fully vaccinated per hundred, and we do find insightful linear correlations.

As a first insight, herd immunity is directly correlated with Total vaccines, people vaccinated and daily vaccinations. As economic factors that influence the progress of herd immunity we find that GDP: Gross domestic product (million current US$) and the capacity to export and import of the country influence herd immunity progress. As a strong correlation, Energy production, primary (Petajoules) , population, and GPD growth influence the capacity of developing herd immunity.

04. Analyzing vaccines progress per country

Once we analyze the overall evolution of all country we can make an analysis of each country aiming to solve the questions of Which vaccine has been more administered in each country? And How has been the rhythm of vaccine administration over time with different vaccine types?

For that, the data visualization proposal proposes 3 different types of graphs:

  • A circle chart  with % of vaccine types
  • A histogram with the evolution of vaccines administrations over time
  • Stacked bars for types of vaccines administered per month.

With that in regard, we can make an idea of vaccination progress work in each country.

05. Predicting Herd immunity

This section aims to respond the question. When will the country reach herd immunity? For predicting herd immunity we have selected the framework Prophet, a procedure for forecasting time series data based on an additive model where non-linear trends are fitted into an autoregressive model. This Framework includes some key capabilities for users to tweak and adjust forecasts. You can use human-interpretable parameters to improve the forecasting by adding some key domain knowledge. When we plot the evolution of herd immunity we clearly can see a type of logistic growth curve (showing a behavior in which immunity starts small but increases exponentially, but then starts to decrease towards an asymptote. Under this setup, there is usually some maximum achievable point: the total herd immunity that the country is able to reach, also known as the cap parameter (carrying capacity), and the forecast should saturate at this point. This parameter is different depending of each illness in a pandemic setup. Prophet offers the possibility to tune this parameter, even within a nature of an increasing sequence. Therefore, the ableness of selecting a logistic growth plus the tunable carrying capacity has been key for selecting prophet framework. Following the results, find above a counter showing the time to 40 people vaccinated per each 100 in Iceland.

Do you want a data-driven solution?

projects scaled