ABSTRACT

Objective:

Describe the development, implementation, and utilization of dashboards for epidemiological analysis through open data research during the COVID-19 pandemic.

Methods:

The dashboards were designed to analyze COVID-19 related public data from various sources, including official government data and social media, at world level. Data processing and cleaning techniques were used to join datasets. We calculated Spearman correlation coefficient between the COVID-like symptoms data of the University of Maryland and Facebook Health research, called COVID Trends and Impacts Survey (CTIS) and the official data of notified COVID-19 cases by the Brazilian Health Ministry.

Results:

The dashboards were successful in predicting the onset of new waves of COVID-19 in Brazil. The data analysis revealed a correlation between the CTIS and the official number of cases the country. This article shows the potential of interactive dashboards as a decision-making tool in the context of public health emergencies, as it was used by the official communication of the Rio Grande do Sul state government.

Conclusion:

The use of dashboards for predicting the spread of COVID-19 in Brazil was a useful tool for decision-making. To anticipate waves of the disease gives time so that these decisions can be potentially more assertive. This drafts the need of more interdisciplinary actions of this nature, with visualization tools on epidemiologic research.

Keywords:
Epidemiology; Decision making; COVID-19; Open data research

INTRODUCTION

The COVID-19 pandemic highlighted the need to improve the availability and visualization of open data in Brazil so that public health decisions could be made quickly and assertively¹1 Rodriguez-Morales AJ, Lopardo G, Verbanaz S, Orduna T, Lloveras S, Azeñas-Burgoa JM, et al. Latin America: situation and preparedness facing the multi-country human monkeypox outbreak. Lancet Reg Health Am. 2022; 13: 100318. https://doi.org/10.1016/j.lana.2022.100318
https://doi.org/10.1016/j.lana.2022.1003... . Many people use data dashboards, sometimes provided by governments, for decision making²2 Peddireddy AS, Xie D, Patil P, Wilson ML, Machi D, Venkatramanan S, et al. From 5Vs to 6Cs: operationalizing epidemic data management with COVID-19 surveillance. medRxiv. 2020: 2020.1.27.20220830. https://doi.org/10.1101/2020.10.27.20220830
https://doi.org/10.1101/2020.10.27.20220... . As SARS-CoV-2 has quickly affected several countries, governments, press professionals, and scientific communicators had to create ways to inform the population in a precise and didactic way, requiring the creation of data dashboards with explanations about the health situation of each municipality, state, and country.

These dashboards have been and still are useful in the current scenario precisely because they standardize different data for visualizing and communicating information³3 Midway SR. Principles of effective data visualization. Patterns. 2020; 1(9): 100141. https://doi.org/10.1016/j.patter.2020.100141
https://doi.org/10.1016/j.patter.2020.10... . Two successful examples in this context are the Our World in Data website⁴4 Our World in Data. Coronavirus data explorer [Internet]. [cited on Apr 16, 2022]. Available at: https://ourworldindata.org/
https://ourworldindata.org/... and the Johns Hopkins University (JHU) dashboard⁵5 Johns Hopkins Coronavirus Resource Center. How to use our data [Internet]. [cited on Apr 23, 2022]. Available at: https://coronavirus.jhu.edu/about/how-to-use-our-data
https://coronavirus.jhu.edu/about/how-to... . Both sites focus, in the context of COVID-19, mainly on information on cases and deaths.

With this in mind, Rede Análise⁶6 Rede Análise. Sobre nós [Internet]. [cited on Sep 08, 2022]. Available at: https://redeanalise.com.br/sobre-nos/
https://redeanalise.com.br/sobre-nos/... , an interdisciplinary collective of researchers and experts, created dashboards with the compilation of multiple open databases from Brazil and the world to increase the assertiveness of decision making. These dashboards contain vaccination coverage, population mobility, hospital data, case and death curves, growth rate, number of people reporting symptoms and the use of masks, in addition to data on Severe Acute Respiratory Syndrome (SARS). All data used in the dashboard is open, that is, available to the general public, which expands research possibilities.

One of the highlights is the COVID-19 Trends and Impact Survey (CTIS), which was carried out jointly by the University of Maryland and Facebook to consult users of the social network on various issues related to the pandemic. The dashboards mentioned in this article used symptom data, where users, if they chose to answer the survey, reported what symptoms they were feeling at that moment. When symptoms of fever, cough, and shortness of breath/difficulty breathing were combined, respondents were considered to be "COVID-like", and this data were then made available for download. This allowed to anticipate the waves of COVID-19 in several Brazilian states from the second wave onward, when tests were carried out more often⁷7 Brasil. Governo Federal. Monitoramento Laboratorial Exames COVID-19 – RT-PCR. Referente aos Laboratórios Centrais em Saúde Pública – Lacens [Internet]. [cited on Jan 21, 2022]. Available at: https://public.tableau.com/views/PAINEL_COVID_IVIS_PBLICO/PAINEL?:embed=y&:showVizHome=no&:host_url=https%3A%2F%2Fpublic.tableau.com%2F&:embed_code_version=3&:tabs=yes&:toolbar=yes&:animate_transition=yes&:display_static_image=no&:display_spinner=no&:display_overlay=yes&:display_count=yes&:language=pt&:loadOrderID=0
https://public.tableau.com/views/PAINEL_... , helping decision makers based on this anticipation.

The dissemination of information from these dashboards was done using the data storytelling technique, which is a relatively new technique in the area of public health. The concept of data storytelling emerged in 2015⁸8 Knaflic CN. Storytelling with data: a data visualization guide for business professionals. Nova Jersey: Wiley; 2015., focusing on business professionals, and involves demonstrating technical information in a way that decision makers can reach more assertive conclusions even without specialized technical knowledge.

Visualization favors the transmission of the message, especially when data is abundant and needs to be translated into information for a wide audience. When telling a story, reader or users are guided to reach the information clearly and objectively. Through interactivity, visual dynamism, and emphasis on referenced data, one hopes for successful communication. The objectives of this work were to describe the development and implementation of dashboards for the epidemiological analysis of open data during the COVID-19 pandemic and to evaluate the correlation between trends evidenced from social media with those recorded by the public health surveillance system.

METHODS

The Rede Análise dashboards were created with the Microsoft Power BI tool⁹9 Microsoft Power BI. Apresentando o Copilot no Power BI [Internet]. [cited on Jan 14, 2023]. Available at: https://powerbi.microsoft.com/pt-br/
https://powerbi.microsoft.com/pt-br/... , powered by public data monitoring the COVID-19¹⁰10 Microsoft Power BI. Rede Análise. Painéis de dados [Internet]. [cited on Apr 15, 2022]. Available at: https://app.powerbi.com/view?r=eyJrIjoiNzI2NTc5NzUtZWJlMC00NGFlLTljOWEtYzNjMGFkODVjM2Y4IiwidCI6IjRkOWRkMWFmLTgzY2UtNGU5Yi1iMDkwLWIwNTQzY2NjMmIzMSIsImMiOjh9
https://app.powerbi.com/view?r=eyJrIjoiN... pandemic. The extraction of these data was done in a non-automated way, using the download options via specific websites mentioned below. Five dashboards were created, as shown in Table 1. For counts of cases, deaths, and hospitalizations in different countries, we used the Our World In Data database, which in turn extracts information from the JHU platform⁵5 Johns Hopkins Coronavirus Resource Center. How to use our data [Internet]. [cited on Apr 23, 2022]. Available at: https://coronavirus.jhu.edu/about/how-to-use-our-data
https://coronavirus.jhu.edu/about/how-to... .

For the dashboard of cases and deaths in Brazil (Figure 1), the database used was that of the Ministry of Health¹¹11 Brasil. Coronavírus. Painel coronavírus [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://covid.saude.gov.br/
https://covid.saude.gov.br/... . The Seade Foundation database was used for hospitalizations in the state of São Paulo (SP)¹²12 Governo do Estado de São Paulo. Seade coronavírus [Internet]. 2022 [cited on Apr 23, 2022]. Available at: https://www.seade.gov.br/coronavirus/#
https://www.seade.gov.br/coronavirus/#... . For hospitalizations in the state of Rio Grande do Sul (RS), data from the RS State Health Department were used¹³13 Governo do Estado do Rio Grande do Sul. Monitoramento COVID-19. Internações hospitalares [Internet]. 2022 [cited on Apr 23, 2022]. Available at: https://covid.saude.rs.gov.br/
https://covid.saude.rs.gov.br/... . Only hospitalization data from RS and SP were used, which are available daily and follow the same structure, allowing for historical analysis. For mobility data, the database used was Google¹⁴14 Google. COVID-19: relatórios de mobilidade da comunidade [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://www.google.com/covid19/mobility/
https://www.google.com/covid19/mobility/... . Regarding symptoms and use of masks, we used the CTIS database¹⁵15 University of Maryland. The University of Maryland Social Data Science Center Global COVID-19 Trends and Impact Survey in partnership with Facebook [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://covidmap.umd.edu/
https://covidmap.umd.edu/... . For SARS data, we used the Influenza Epidemiological Surveillance System (Sistema de Vigilância Epidemiológica da Gripe – SIVEP-GRIPE) database, obtained through OpenDataSUS; finally, for vaccination, data from the National Immunization Plan Information System (Sistema de Informação do Plano Nacional de Imunizações – SI-PNI) were used¹⁶16 Brasil. Ministério da Saúde. DataSUS. SRAG 2021 e 2022 – Banco de Dados de Síndrome Respiratória Aguda Grave – incluindo dados da COVID-19 [Internet]. 2022 [cited on Apr 23, 2022]. Available at: https://opendatasus.saude.gov.br/dataset/srag-2021-e-2022
https://opendatasus.saude.gov.br/dataset... –¹⁸18 Brasil. Ministério da Saúde. DataSUS. Campanha Nacional de Vacinação contra Covid-19 [Internet]. 2022 [cited on Apr 23, 2022]. Available at: https://opendatasus.saude.gov.br/dataset/covid-19-vacinacao
https://opendatasus.saude.gov.br/dataset... .

Procedures for data extraction

Two types of files were extracted from the Our World In Data database. The first file is full_data.csv¹⁹19 covid-19-data/full_data.csv at master owid/covid-19-data [Internet]. GitHub; 2022 [cited on Apr 23, 2022]. Available at: https://github.com/owid/covid-19-data/blob/master/public/data/jhu/full_data.csv
https://github.com/owid/covid-19-data/bl... which is obtained from the public domain github of Our World in Data, which in turn is powered by JHU. This file contains new cases, new deaths, total cases, total deaths, weekly cases, weekly deaths, fortnightly cases and fortnightly deaths reported by date, as of 01/22/2020, for 216 countries²⁰20 owid/covid-19-data. COVID-19 dataset by Our World in Data [Internet]. GitHub; 2022. [cited on Apr 15, 2022]. Available at: https://github.com/owid/covid-19-data
https://github.com/owid/covid-19-data... . In addition to countries, there are also regions of interest, such as "World", "Europe", "Low-income countries", among others.

The second file is owid-covid-data.csv, which contains, for the same 228 locations, in addition to the data already contained in full_data.csv, the data dictionary that can be found in the supplementary material. We highlight the seven-day moving average columns for new cases and deaths and per million inhabitants²¹21 Google Docs. Dicionario_de_dados_owid_rede [Internet]. 2022. [cited on Dec 12, 2022]. Available at: https://docs.google.com/spreadsheets/d/1CfPOj6HXALblPcUJ1Bze2hLnyElkEiFsx3bHY9VQam0/edit?usp=sharing&usp=embed_facebook
https://docs.google.com/spreadsheets/d/1... . This file was extracted from the public domain data on Our World in Data’s github.

The full_data.csv file is then imported using the Python programming language, with a script that calculates the growth rates of COVID-19 cases and deaths, among other indicators. The script and field table can be found in the supplementary material. The growth rate formula, used both in cases and deaths, is x = (n2/n1) -1, where x is the growth rate, n2 is the number of cases/deaths on a date, and n1 is the number of cases/deaths from the immediately previous day. This script also calculates the seven-day moving average of new cases reported per day, as well as new deaths reported per day, with the purpose of smoothing the graphs, as notifications drop during weekends and holidays. The moving average of new cases reported per day is calculated by adding the new cases in the last seven days and dividing by seven, generating a new variable in the database, with the same calculation being made with the new deaths reported per day. The owid-covid-data.csv source also has data on doses of the COVID-19 vaccine administered per day, in addition to the estimated population of each country. With this, the script also calculates vaccination coverage to make available on the dashboards.

The extraction of data on cases, deaths from Brazil, its states and municipalities is done by transferring a file, obtained from the Ministry of Health’s official COVID-19 monitoring platform¹¹11 Brasil. Coronavírus. Painel coronavírus [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://covid.saude.gov.br/
https://covid.saude.gov.br/... . The data are in a compressed file that contains several separate CSV files. Data processing is done by executing a script programmed in python language, which can be found in the supplementary material, as well as in the fields table.

The extraction of hospitalization data from the state of São Paulo is done through the file plano_sp_leitos_internacoes_serie_nova_variacao_semanal.csv, without any extra calculations. The extraction of hospitalization data from the state of Rio Grande do Sul is done using the file transparencia_dados_covid.csv, without any extra calculations.

Mobility data is extracted using the Region_Mobility_Report_CSVs.zip file, which contains 405 files, three files for each of the 135 countries. The three files refer to the years 2020, 2021, and 2022. They are aggregated into one for the countries used in the dashboards (currently Brazil, Chile, Uruguay, United States, Israel, United Kingdom, Indonesia, and India).

The extraction of symptom data is done with a script in python language that fetches the data from the Application Program Interface (API) of the University of Maryland¹⁵15 University of Maryland. The University of Maryland Social Data Science Center Global COVID-19 Trends and Impact Survey in partnership with Facebook [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://covidmap.umd.edu/
https://covidmap.umd.edu/... ,²²22 Loezer L. COVID-19 World Survey Data API [Internet]. GitHub; 2021 [cited on Apr 23, 2022]. Available at: https://github.com/loezerl/covid19-worldsurverydata-brazil/blob/c5398be80420e38c6ede1a748e1b54333ce403e1/report.py
https://github.com/loezerl/covid19-world... . This script allows choosing which indicator to be used, as well as for which states or regions of Brazil. The indicators covid (people who reported feeling COVID-19-like symptoms), flu (people who reported feeling influenza-like symptoms), and mask (people who reported wearing a mask when leaving home) were used. The extraction of SARS data is done through separate files, for the years 2020, 2021, and 2022, without additional calculations. The extraction of vaccination data is done through separate files, one for each state in Brazil²³23 Cota W. Vaccination data [Internet]. GitHub; 2022 [cited on Sep 08, 2022]. Available at: https://github.com/wcota/covid19br-vac
https://github.com/wcota/covid19br-vac... .

All files are used as data sources in the Microsoft Power BI Desktop software, to assemble visualization dashboards. These fonts are arranged in all dashboards, in the lower right corner. Power BI was used due to its low implementation cost and also the agility in assembling the dashboards.

It is important to mention that there are typographical errors in database notifications. To emphasize transparency and replicability in the use of data, we chose to keep these errors in the dashboards, explaining them to help educate the population, demonstrating how important it is to keep the database correct and updated. In addition to the educational process of highlighting the need for data curation, maintaining inconsistencies allows analyses to be replicated more reliably by different actors.

These data are presented through data storytelling, with the objective of accelerating understanding on the part of decision makers who, in many cases, do not have similar technical knowledge to those who carried out the analysis and/or synthesis of the information⁸8 Knaflic CN. Storytelling with data: a data visualization guide for business professionals. Nova Jersey: Wiley; 2015.. Thus, it is possible to deliver data both on social networks, to individual decision makers — such as heads of families — and to public and private managers, who need to make their decisions quickly and assertively. This data delivery is constantly updated as data sources are updated, so that decision makers can always search for data dynamically. The script that performs the calculations reported in this section is available on Rede Análise’s GitHub²⁴24 RedeAnalise-Covid-19/painesdedados. Paineis de dados – dashboards [Internet]. GitHub; 2023 [cited on Sep 02, 2023]. Available at: https://github.com/RedeAnalise-Covid-19/paineisdedados
https://github.com/RedeAnalise-Covid-19/... .

RESULTS

Mobility and Symptoms dashboards were able to demonstrate in advance the beginning of the second and third waves of COVID-19 in several states in Brazil (Figure 2), through the analysis of data from the CTIS survey. During the first wave, the data were not yet consolidated, with its publication ending in June 2022, making it no longer possible to act in subsequent waves. Furthermore, Hospitalization dashboards also demonstrated the volatility of the ICU occupancy indicator (Figure 3) with the data made available by the state of SP.

The CTIS survey, promoted by the University of Maryland in conjunction with Facebook, worked as follows: users of the social network are invited daily to answer a questionnaire where, among various topics, collects possible symptoms that respondents are feeling at that moment¹⁵15 University of Maryland. The University of Maryland Social Data Science Center Global COVID-19 Trends and Impact Survey in partnership with Facebook [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://covidmap.umd.edu/
https://covidmap.umd.edu/... . When symptoms are fever, cough, and shortness of breath/difficulty breathing, respondents are marked as one who may have COVID-19. Dashboards cross-reference this data with case reports by the Ministry of Health in Brazil to show how the research anticipated the rises and falls of the COVID-19 case curve¹⁵15 University of Maryland. The University of Maryland Social Data Science Center Global COVID-19 Trends and Impact Survey in partnership with Facebook [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://covidmap.umd.edu/
https://covidmap.umd.edu/... .

This anticipation can be seen when calculating the Spearman correlation coefficient, which was done in a script in the R language, available in the supplementary material. Table 2 presents the correlation coefficients for four Brazilian states that had a significant number of respondents, and in two situations:

1. Maintaining the original dates, with research demonstrating the increase in cases approximately 20 days before official notifications of COVID-19 cases.

2. Artificially aligning the dates, that is, delaying the research data by 20 days so that it is aligned with official notifications of COVID-19 cases.

Table 2 shows a significant correlation between symptom survey data and official case notifications released by the Ministry of Health.

In addition to anticipating waves, these dashboards allow seeing and understanding mobility (Figure 4) in Brazil, in its states and municipalities. The dashboard shows data coming from the Community Mobility Reports tool, made available by Google, which, anonymously, collects the number of people visiting six categories of locations: homes, workplaces, public transport stations, parks, markets/pharmacies, and commerce/leisure. These data show how the population’s behavior changed over time during the pandemic¹⁴14 Google. COVID-19: relatórios de mobilidade da comunidade [Internet]. 2022 [cited on Apr 15, 2022]. Available at: https://www.google.com/covid19/mobility/
https://www.google.com/covid19/mobility/... .

In addition to these specific data, the dashboards also show data on cases and deaths together with the growth rate of the two indicators, from all states and municipalities in Brazil, in addition to 228 locations worldwide. The data show, in addition to the history of the entire pandemic period, the trend of the last 20 days and the behavior of the growth rate of cases and deaths. All are indicators that, when analyzed together and explained through data storytelling, can assist in the decision-making process²⁵25 Isaac Schrarstzhaupt’s Threads – Thread Reader App [Internet]. [cited on Apr 15, 2022]. Available at: https://threadreaderapp.com/user/schrarstzhaupt
https://threadreaderapp.com/user/schrars... .

The dashboards also serve as a window to understand data on SARS cases and deaths reported in SIVEP-GRIPE, with filters by state, municipality, age range, race/color and cause of SARS. This data helps to understand how each age group suffers the impacts of the disease²⁶26 Duarte M, Mariz R. Uma criança morreu a cada três dias por Covid enquanto país debatia vacinação infantil, mostra levantamento do GLOBO. Jornal O Globo [Internet]. [cited on Apr 15, 2022]. Available at: https://oglobo.globo.com/saude/uma-crianca-morreu-cada-tres-dias-por-covid-enquanto-pais-debatia-vacinacao-infantil-mostra-levantamento-do-globo-1-25394548
https://oglobo.globo.com/saude/uma-crian... .

The dashboards were made available to the population mainly through threads on the social network Twitter²⁵25 Isaac Schrarstzhaupt’s Threads – Thread Reader App [Internet]. [cited on Apr 15, 2022]. Available at: https://threadreaderapp.com/user/schrarstzhaupt
https://threadreaderapp.com/user/schrars... , which reached a total of 74.9 million impressions (views) between January 2021 and January 2022. In May 2021, the dashboards also began to be published in interactive links thanks to the support of Doctors Without Borders. This support allowed the dashboards to be available through reduced links on the bit.ly platform, where there were 25,934 engagements with these links. In addition to the dashboards themselves, the storytelling used to demonstrate the data on social networks, in the analyses available on the Rede Análise²⁷27 Rede Análise. Terceira onda de covid-19 de 2022 [Internet]. 2022 [cited on May 31, 2023]. Available at: https://redeanalise.com.br/brasil/terceira-onda-de-covid-19-de-2022/
https://redeanalise.com.br/brasil/tercei... website and in live streams is another factor to help in the assertiveness of decisions, as it supports both individual decision makers and public health decision makers²⁸28 You Tube. Live: atualização das ações para combate à pandemia no RS [Internet]. 2021 [cited on Apr 15, 2022]. Available at: https://youtu.be/cNgGqjBqoyo?t=859
https://youtu.be/cNgGqjBqoyo?t=859... .

DISCUSSION

The use of data dashboards gained notoriety with the COVID-19 pandemic, due to their easy navigation. However, the dashboards themselves cannot explain what is happening and what could happen in current and future scenarios, requiring data analysis/interpretation to extract this information. To ensure easier navigation and increased accessibility, the dashboards were communicated through data storytelling, on social media, with a wide reach of the population and public managers²⁸28 You Tube. Live: atualização das ações para combate à pandemia no RS [Internet]. 2021 [cited on Apr 15, 2022]. Available at: https://youtu.be/cNgGqjBqoyo?t=859
https://youtu.be/cNgGqjBqoyo?t=859... . Furthermore, these dashboards can be accompanied by a "troubleshooting"’ containing frequently asked questions, to reduce the difficulty in understanding the graphics, since this understanding does not only depend on what is in the graph itself, but also on the objective and knowledge of who is seeking the information.

Based on this, these dashboards were created to, through data storytelling, assist in the population’s decision-making, associating social networks in the dissemination of information relevant to the current scenario. With this method, it was possible to visually demonstrate the alert situations that occurred during the pandemic¹⁰10 Microsoft Power BI. Rede Análise. Painéis de dados [Internet]. [cited on Apr 15, 2022]. Available at: https://app.powerbi.com/view?r=eyJrIjoiNzI2NTc5NzUtZWJlMC00NGFlLTljOWEtYzNjMGFkODVjM2Y4IiwidCI6IjRkOWRkMWFmLTgzY2UtNGU5Yi1iMDkwLWIwNTQzY2NjMmIzMSIsImMiOjh9
https://app.powerbi.com/view?r=eyJrIjoiN... ,²⁵25 Isaac Schrarstzhaupt’s Threads – Thread Reader App [Internet]. [cited on Apr 15, 2022]. Available at: https://threadreaderapp.com/user/schrarstzhaupt
https://threadreaderapp.com/user/schrars... ,²⁹29 Carnevalli S. Data Storytelling Canvas: conte a história do seus dados [Internet]. 2020 [cited on Apr 15, 2022]. Available at: http://datastorytelling.com.br/data-storytelling-canvas-conte-a-historia-do-seus-dados/
http://datastorytelling.com.br/data-stor... .

In addition to the navigation that can be done by anyone, the dissemination of data through the sum of dashboards with the data storytelling technique is of great value to decision makers. In many situations, the data is the responsibility of a technical area, and decision making comes from managers who do not have the same level of knowledge³⁰30 Segel E, Heer J. Narrative visualization: telling stories with data. IEEE Transactions on Visualization and Computer Graphics. 2010; 16(6): 1139-48.. Through data storytelling, a broader understanding of the data can be achieved by decision makers, who can then be more assertive in their choices. One can see the use of storytelling in communication on social networks (Figure 5) and in its use by decision makers themselves²⁸28 You Tube. Live: atualização das ações para combate à pandemia no RS [Internet]. 2021 [cited on Apr 15, 2022]. Available at: https://youtu.be/cNgGqjBqoyo?t=859
https://youtu.be/cNgGqjBqoyo?t=859... .

Both public managers, citizens, and the press³¹31 Duarte M. Covid-19: Governo não compra vacinas e já faltam doses a crianças de 3 e 4 anos. O Globo [Internet]. 2022 [cited on Jan 04, 2023]. Available at: https://oglobo.globo.com/saude/noticia/2022/08/covid-19-governo-nao-compra-vacinas-e-ja-faltam-doses-a-criancas-de-3-e-4-anos.ghtml
https://oglobo.globo.com/saude/noticia/2... can use and navigate the dashboards to improve their decision making, which confirms the importance of having an increasingly widespread culture of open data³²32 McKiernan EC, Bourne PE, Brown CT, Buck S, Kenall A, Lin J, et al. How open science helps researchers succeed. eLife. 2016; 5: e16800. https://doi.org/10.7554/eLife.16800
https://doi.org/10.7554/eLife.16800... . The state of Rio Grande do Sul used the symptom data dashboard from the CTIS survey as decision-making data due to the anticipation of cases, demonstrated in the methods. As testing has delay problems³³33 Torres I, Sippy R, Sacoto F. Assessing critical gaps in COVID-19 testing capacity: the case of delayed results in Ecuador. BMC Public Health. 2021; 21(1): 637. https://doi.org/10.1186/s12889-021-10715-x
https://doi.org/10.1186/s12889-021-10715... , the case indicator ends up having a delay in its notification, and indicators that anticipate this wave and are easily accessible, such as in an online dashboard, are of great contribution.

The most acute period of the pandemic brought the public health context closer to the population, and science communication, when becoming a protagonist, needed assertiveness and accuracy in information³⁴34 Matta G. Science communication as a preventative tool in the COVID19 pandemic. Humanit Soc Sci Commun. 2020; 7(1): 159. https://doi.org/10.1057/s41599-020-00645-1
https://doi.org/10.1057/s41599-020-00645... . Data storytelling appears as an ally, because, by creating analogies together with data dashboards, it is possible to illustrate situations to the point where understanding is achieved more effectively³⁵35 YouTube. Como saber se os casos/óbitos de COVID-19 irão piorar? [Internet]. 2022 [cited on Dec 13, 2022]. Available at: https://www.youtube.com/
https://www.youtube.com/... .

There are practical examples of data storytelling in several publications made during the pandemic health emergency²⁷27 Rede Análise. Terceira onda de covid-19 de 2022 [Internet]. 2022 [cited on May 31, 2023]. Available at: https://redeanalise.com.br/brasil/terceira-onda-de-covid-19-de-2022/
https://redeanalise.com.br/brasil/tercei... ,³⁶36 Rede Análise. Aumento de internações no interior do estado de São Paulo entre as festas de fim de ano [Internet]. 2022 [cited on May 31, 2023]. Available at: https://redeanalise.com.br/covid-19/aumento-de-internacoes-no-interior-do-estado-de-sao-paulo-entre-as-festas-de-fim-de-ano/
https://redeanalise.com.br/covid-19/aume... where the narrative flow of storytelling is used, which consists of telling the audience what will be demonstrated, in chronological order and, at the end, summarize it, with the most important points and also with the points where there are still gaps in data⁸8 Knaflic CN. Storytelling with data: a data visualization guide for business professionals. Nova Jersey: Wiley; 2015.. A limitation of this work is that we do not have a formal qualitative assessment of dashboard users, such as focal interviews or evaluative questionnaires. However, storytelling received positive support from the feedback from Twitter users, where the threads reached more than 70 million views, a quantitative metric (dependent on audience interaction to generate absolute value) in a defined space of time that reflects the scope of data being transformed into information.

Several similar initiatives have been observed during the COVID-19 pandemic health emergency, both governmental and by volunteer researchers, who played a very important role in disseminating data. In Brazil, there are independent initiatives such as Brasil.Io³⁷37 Brasil.IO. O brasil em dados libertos; Repositório de dados públicos disponibilizados em format acessível. Especial COVID-19: dados por município [Internet]. 2023 [cited on Aug 26, 2023]. Available at: https://brasil.io/home/
https://brasil.io/home/... , which brings data and dashboards, and governmental ones, such as InfoGripe³⁸38 Bastos LS, Economou T, Gomes MFC, Villela DAM, Coelho FC, Cruz OG, et al. A modelling approach for correcting reporting delays in disease surveillance data. Stat Med. 2019; 38(22): 4363-77. https://doi.org/10.1002/sim.8303
https://doi.org/10.1002/sim.8303... , from Fiocruz, which generated a model to correct the delay in reporting data on hospitalizations due to SARS, which is very important for the decision making. In South Africa, there is the South Africa Medical Research Council (SAMRC)³⁹39 South African Medical Research Council. Report on weekly deaths in South Africa [Internet]. 2023 [cited on Aug 26, 2023]. Available at: https://www.samrc.ac.za/research-reports/report-weekly-deaths-south-africa
https://www.samrc.ac.za/research-reports... , with the tracking of deaths due to COVID-19, and also initiatives by researchers to understand the excess of deaths from all causes⁴⁰40 Karlinsky A, Kobak D. Tracking excess mortality across countries during the COVID-19 pandemic with the World Mortality Dataset. eLife. 2021; 10: e69336. https://doi.org/10.7554/eLife.69336
https://doi.org/10.7554/eLife.69336... ,⁴¹41 Governo do Estado do Rio Grande do Sul. Secretaria de Panejamento, Governança e Gestão. Departamento de Economia e Estatística. Boletins Coronavírus [Internet]. [cited on Sep 06, 2023]. Available at: https://boletinscoronavirus.dee.rs.gov.br/
https://boletinscoronavirus.dee.rs.gov.b... .

However, there are still limitations in terms of notification and the availability of higher quality and more frequent public data, there are no data regarding the date of onset of symptoms, date of death, vaccination status, previous infections, among others. We could have them, always respecting the principles of anonymization. This will further improve the creation and use of dashboards; Therefore, storytelling will further assist in individual decision making and in the formulation of public policies.

In the coming years, due to the post-COVID⁴²42 Heine J, Schwichtenberg K, Hartung TJ, Rekers S, Chien C, Boesl F, et al. Structural brain changes in patients with post-COVID fatigue: a prospective observational study. EClinicalMedicine. 2023; 58: 101874. https://doi.org/10.1016/j.eclinm.2023.101874
https://doi.org/10.1016/j.eclinm.2023.10... characteristics and the large number of cases, it is possible that health systems will be pressured by the increase in patients with sequelae arising from this period. Knowing what these consequences are, their frequency and incidence will help both the public authorities to provide funds and the general public to be informed, seeking faster assistance. The democratization of information through constant availability of public data reaches more people and, as it is a free tool for those who use it, it has enormous potential to support public health policies. Dashboards like these tend to be increasingly common, with various applications, such as the monitoring of other respiratory pathogens carried out by Instituto Todos pela Saúde⁴³43 Instituto Todos pela Saúde. Monitoramento de patógenos respiratórios – relatório 21 [Internet]. 2023 [cited on Aug 26, 2023]. Available at: https://www.itps.org.br/pesquisa-detalhe/monitoramento-de-patogenos-respiratorios-relatorio-21
https://www.itps.org.br/pesquisa-detalhe... .

ACKNOWLEDGMENTS:

We would like to thank Doctors Without Borders – MSF, for their support in publishing the dashboards.

REFERENCES

Publication Dates

History