Online dashboard and data analysis approach for assessing COVID-19 case and death data

Hector Florez; Sweta Singh

doi:10.12688/f1000research.24164.1

Home Browse Online dashboard and data analysis approach for assessing COVID-19...

ALL Metrics

-

Views

-

Downloads

Get PDF

Get XML

Export

▬

✚

Software Tool Article

Online dashboard and data analysis approach for assessing COVID-19 case and death data

[version 1; peer review: 2 approved, 1 approved with reservations]

Hector Florez ¹, Sweta Singh²

PUBLISHED 08 Jun 2020

Author details Author details

¹ Universidad Distrital Francisco Jose de Caldas, Bogota, Colombia
² Savitribai Phule Pune University, Pune, India

Hector Florez
Roles: Conceptualization, Formal Analysis, Software, Writing – Original Draft Preparation

Sweta Singh
Roles: Conceptualization, Formal Analysis, Validation, Writing – Review & Editing

OPEN PEER REVIEW

REVIEWER STATUS

This article is included in the Emerging Diseases and Outbreaks gateway.

This article is included in the Coronavirus collection.

Abstract

The 2019-2020 global pandemic has been caused by a disease called coronavirus disease 2019 (COVID-19). This disease has been caused by the Severe Acute Respiratory Syndrome coronavirus-2 (SARS-CoV-2). By April 30 2020, the World Health Organization reported 3,096,626 cases and 217,896 deaths, which implies an exponential growth for infection and deaths worldwide. Currently, there are various computer-based approaches that present COVID-19 data through different types of charts, which is very useful to recognise its behavior and trends. Nevertheless, such approaches do not allow for observation of any projection regarding confirmed cases and deaths, which would be useful to understand the trends of COVID-19. In this work, we have designed and developed an online dashboard that presents actual information about COVID-19. Furthermore, based on this information, we have designed a mathematical model in order to make projections about the evolution of cases and deaths worldwide and by country.

Keywords

COVID-19, SARS-CoV-2, Data analysis, Mathematical model

Corresponding author: Hector Florez

Competing interests: No competing interests were disclosed.

Grant information: The author(s) declared that no grants were involved in supporting this work.

Copyright: © 2020 Florez H and Singh S. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

How to cite: Florez H and Singh S. Online dashboard and data analysis approach for assessing COVID-19 case and death data [version 1; peer review: 2 approved, 1 approved with reservations]. F1000Research 2020, 9:570 (https://doi.org/10.12688/f1000research.24164.1) First published: 08 Jun 2020, 9:570 (https://doi.org/10.12688/f1000research.24164.1) Latest published: 08 Jun 2020, 9:570 (https://doi.org/10.12688/f1000research.24164.1)

Introduction

Coronavirus disease 2019 (COVID-19) is caused by Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2), which is a virus strain that causes respiratory illness. This virus was identified in December, 2019 and its first infection case was reported on December 30, 2019 in Wuhan city located in Hubei providence, China¹. The World Health Organization (WHO) recognized the disease as a pandemic on March 11 2020². The WHO have reported the following numbers of cases and deaths worldwide: January 31, 9,847 cases and 213 deaths³; February 29, 85,961 cases and 2,941 deaths⁴; March 31, 754,933 cases and 36,522 deaths⁵; and April 30, 3,096,626 cases and 217,896 deaths⁶. These reports show an increment of cases in February of 872.97%, in March of 878.22%, and in April of 410.18%, as well as an increment of deaths in February of 1,387.26%, in March of 1,241.82%, and in April of 596.61%⁷. Based on these percentiles, we can observe an exponential growth of the disease.

Based on this previous data, it is important to analyze the evolution of the disease in order to make decisions that tackle the growth rate of cases and deaths. Therefore, in this work, we have designed and developed an online dashboard which presents two types of results: (1) actual information about the COVID-19 using different deployment strategies; and (2) based on the actual information, a projection in order to estimate the future behavior of the disease calculated worldwide and by country. To achieve this projection, we have designed a mathematical model based on previous data reported by the WHO.

Methods

Implementation

To analyze the current state of COVID-19, we have created an online information system called COVID-19 Dashboard, which is accessible from: https://covid19.itiud.org/. For the design of COVID-19 Dashboard, Unified Modeling Language (UML)⁸ was used allowing the description of the software based on the Object Oriented Paradigm^9,10. It has been developed using: a) PHP 7.4.5 (https://www.php.net/) supported by Apache 2.4.43 (https://www.apache.org/); b) Bootstrap 4.4.1 (https://www.getbootstrap.com/) for responsive front-end; c) jQuery 3.3.1 (https://www.jquery.com/) as JavaScript library.

COVID-19 Dashboard presents full information about COVID-19 focusing on cases and deaths worldwide and by country. In addition, it includes a mathematical model in order to calculate projections for future cases and deaths. The data source used in COVID-19 Dashboard is taken from the WHO (https://covid19.who.int/) and World Population Review (https://www.worldpopulationreview.com/). The first data set is accessed by COVID-19 Dashboard through the URL https://dashboards-dev.sprinklr.com/data/9043/global-covid19-who-gis.json. This information is in JSON format; then, the system includes the algorithms to decode this format before processing the information. The second data set is downloaded in CSV format and included in the persistence layer of the system.This project deploys different graphics to analyze the information from different perspectives¹¹. These graphics are generated based on the JavaScript library Chartkick 2.7.2 (https://www.chartkick.com/), which is supported by Google Charts (https://developers.google.com/chart).

To perform the projection for estimating future cases and deaths, we developed a mathematical model focused on a quadratic equation¹² based on data of present day and 30 previous days. With the quadratic equation, projected cases and deaths for the next days up to 90 days are calculated. The quadratic equation obtained for each projection is deployed supported by the JavaScript library Mathjax (https://www.mathjax.org/).

Figure 1 presents the components of the system, where all mentioned libraries are connected to COVID-19 Dashboard. The conceptual models of COVID-19 Dashboard have been created using the modeling tool UML Designer 9.0.0 (http://www.umldesigner.org/). The selected Integrated Development Environment (IDE) was Eclipse-PHP 2020-03 (https://www.eclipse.org/) and the server used to host the project was Debian Linux Server 10.4 (https://www.debian.org/)¹³.

Figure 1. COVID-19 Dashboard components.

Modeling projections for cases and deaths

For the mathematical model, let three points P₁ = (x₁, y₁), P₂ = (x₂, y₂), and P₃ = (x₃, y₃) such that y₁ corresponds to number of cases or deaths of the date 30 days prior to the present day, x₁ = 0, y₂ corresponds to number of cases or deaths of the date 15 days prior to the present day, x₂ = 15, y₃ corresponds to number of cases or deaths of the present day, and x₁ = 30. Based on the quadratic equation defined as f (x) = ax² + bx + c, let the equation system presented in (1), (2), and (3).

y_{1} = a x_{1}^{2} + b x_{1} + c (1)

y_{2} = a x_{2}^{2} + b x_{2} + c (2)

y_{3} = a x_{3}^{2} + b x_{} + c (3)

Since x₁ = 0 from (1):

y_{1} = c (4)

Subtracting (2) from (4):

y_{1} - y_{2} = - a x_{2}^{2} - b x_{2} (5)

Isolating the variable b from (5):

b = \frac{y_{1} - y_{2} + a x_{2}^{2}}{- x_{2}} (6)

Subtracting (2) from (3):

y_{3} - y_{2} = a (x_{3}^{2} - x_{2}^{2}) + b (x_{3} - x_{2}) (7)

Replacing b in (7):

\begin{array}{l} y_{3} - y_{2} = a (x_{3}^{2} - x_{2}^{2}) + \frac{y_{1} - y_{2} + a x_{2}^{2}}{- x_{2}} (x_{3} - x_{2}) \\ y_{3} - y_{2} = a (x_{3}^{2} - x_{2}^{2}) + \frac{(y_{1} - y_{2}) (x_{3} - x_{2})}{- x_{2}} - a x_{2} (x_{3} - x_{2}) \\ y_{3} - y_{2} = a ((x_{3}^{2} - x_{2}^{2}) - x_{2} (x_{3} - x_{2})) - \frac{(y_{1} - y_{2}) (x_{3} - x_{2})}{x_{2}} (8) \end{array}

Isolating the variable a from (8):

a = \frac{y_{3} - y_{2}}{((x_{3}^{2} - x_{2}^{2}) - x_{2} (x_{3} - x_{2}))} + \frac{(y_{1} - y_{2}) (x_{3} - x_{2})}{x_{2} ((x_{3}^{2} - x_{2}^{2}) - x_{2} (x_{3} - x_{2}))} (9)

Finally, isolating the variable c from (4):

c = y_{1} (10)

Therefore, based on a, b, and c applied to the quadratic equation, the projected cases and deaths are calculated. In addition, calculated quadratic equations will be adjusted when new information is reported by the WHO. Furthermore, COVID-19 Dashboard offers a projection for 90 days; nevertheless, for some countries, the quadratic equation could rise to a maximum point; in these cases, the algorithms will stop calculating and thus the projected days will be lower than 90 days.

Consequently, COVID-19 Dashboard allows analyzing confirmed cases, deaths, and mortality rates as well as projecting cases and death deploying information in: a) geographic charts to understand the country distribution of the disease, b) bar charts to recognize the most impacted countries, c) column charts to present the evolution of the disease over time, and d) line charts to estimate the behavior of the disease.

Operation

COVID-19 Dashboard is an online responsive system; therefore, it is accessible for users via a web browser through desktops, laptops, smart phones, and tablets. Its front-end can automatically adjust to any screen resolution in order to deploy the information in the best possible way.

Results

COVID-19 Dashboard is able to process the data provided by the WHO related to COVID-19 rates and World Population Review related to the current world population. This system deploys up-to-date information as follows:

Cases and deaths by country. This information is presented in two ways: geographic charts and bar charts in order to recognize the most impacted countries.
Cases and deaths by country per million people. This information is also presented by geographic and bar charts.
Mortality rate by country. For this information, the mortality rate is calculated for each country and is presented by geographic and bar charts.
Cases and deaths by date worldwide and by selected country. This information is calculated based on the daily information reported by each country. This is presented in two column charts. The first column chart presents the daily cases, while the second column chart presents the cumulative cases.
Projected cases and deaths worldwide and by selected country. This information is calculated using the quadratic equation obtained through the developed mathematical model (as above). It is presented using a line chart with two series. The first presents the actual cases or deaths for the last 31 days. The second presents the projected cases or deaths up to 90 days after the last day reported.

As mentioned above, the number of cases, deaths, and mortality rates are deployed using geographic charts. Figure 2 presents a geographic chart with the number of deaths by country, where red implies a large number of deaths and light blue implies a low number of deaths. On the date that this chart was obtained from COVID-19 Dashboard, the United States had the largest number of deaths worldwide followed by Spain, France, Italy and the United Kingdom. However, just using the color coding, it is not clear if Spain, France, Italy or the United Kingdom has the second highest number of deaths. Similarly, for the lower number of deaths seen in Belgium, Brazil, Germany, Canada, Turkey, Netherlands, Iran, and China, it is not clear which has the lowest number of deaths using the color coding.

Figure 2. Geographic chart from COVID-19 Dashboard showing the number of deaths by country using colour coding.

Red = high number of deaths; light blue = low number of deaths. Chart captured on May 29^th 2020.

Consequently, it is necessary to sort the countries based on the actual number of deaths in addition to the color coding. Figure 3 presents a fragment of a bar chart with the number of deaths per country sorted by number of deaths. This chart shows that the United States has more than the double of deaths of Italy and the United Kingdom. In addition, Spain and France have almost the same number, but a little less than Italy and United Kingdom, followed by Brazil, while Belgium, Mexico, Germany, Iran, and Canada have similar number of deaths but much less than Spain and France. Finally, Netherlands, India, China, Turkey, Russia, and Sweden have less but a still important number of cases.