Welcome to the CovidR Contest!

eRum2020::CovidR is an open-source contest and pre-conference event launched by the 2020 European R Users Meeting (e-Rum2020), featuring any work done with R around the topic of the COVID-19 pandemic.

The pre-conference virtual event will take place on May 29th, when selected participants will be invited to present their work and the contest winners will be announced and awarded with e-prizes. Accepted contributions (submitted before May 22nd) are made available in this website, where the wider community is invited to provide feedback. The winner(s) will be announced during the conference pre-event on May 29th. Attendance to the event is subject to a free registration, closing on May 22nd.

The set of accepted submissions feature the contributions below. The corresponding gallery pages allow the wider community to browse through and like contributions.


The winners of the CovidR contest, invited to present at the main e-Rum2020 conference are

COVOID: Modelling COVID-19 Transmission and Interventions by Oisin Fitzgerald, Mark Hanly, Tim Churches (UNSW Medicine)
Abstract — COVOID is an evolving but fully-functional R package and accompanying Shiny app for simulation modelling of both COVID-19 transmission and the interventions intended to reduce that transmission, using deterministic compartmental models (DCMs). The built-in Shiny app enables ease of use and demonstration of key concepts to those without R programming backgrounds. The package contains an expanding API for simulating and estimating homogeneous and age-structured SIR, SEIR and extended models. In particular COVOID allows the simultaneous simulation of age specific (e.g. school closures) and general interventions over varying time intervals. This is informed through incorporation of publicly available data on population demographics from the United Nations, age and setting specific contact rates from previously published research and COVID-19 incidence counts from the European CDC. Technical background documentation and vignettes are provided, illustrating the use of the package, including the extended model. Work is ongoing to add the ability to model imported cases, enhancing parameter fitting to observed data, and to expand the Shiny UI to address the extended model and to manage parameter sensitivity analyses and comparisons of intervention scenario.
COVID-19 Data Hub by Emanuele Guidotti, David Ardia
Abstract — Built with R, available in any language, the Hub provides a worldwide, fine-grained, unified dataset helpful for a better understanding of COVID-19. The user can instantly download up-to-date, structured, historical daily data across several official sources. The data are hourly crunched by the R package COVID19 and made available in csv format on a cloud storage, so to be easily accessible from Excel, R, Python… and any other software. We welcome external contributors to join and extend the number of supporting data sources. All sources are properly documented, along with their citation. COVID-19 Data Hub can spot misalignments between data-sources and automatically inform authorities of possible errors. All logs are available at https://covid19datahub.io Available on CRAN.

The following are the contributions awarded at the e-Rum2020 pre-conference event

The Ten-Day Coronavirus Forecast by Ben Phillips, Melissa Makin, Uli Felzmann, Ian Marschner
Abstract — Motivated by genuine concern that many countries were being caught flat-footed by the coronavirus pandemic, we built a shiny app that, for every country, reports estimated case detection and growth rates, and which provides a ten-day projection of active case numbers. The idea was to move us past the paralysis of daily shock and towards important action, but the site also gives us feedback on how our actions are working. The app uses a combination of methods ranging from simple heuristics to sophisticated deconvolution algorithms. State-level reporting is also available for some countries. The data are updated daily, and come from datasets collated by teams at John Hopkins University and Panjab University.
Interactive Dashboard for Real-Time Recording of COVID-19 Outbreak in Japan by Wei Su
Abstract — The project is a website for real-time visualization of the COVID-19 epidemic in Japan, developed mainly using the R language with shiny and other open-source packages. It mainly shows various indicators including, but not limited to, PCR test, positive confirmed, hospital discharge and death, as well as trends in each prefecture in Japan, and there are also a variety of charts such as cluster network, new confirmed case in log scale for users' reference.
Daily Economic Indicators Based on Google Searches by Angelica Becerra et al., trendecon team
Abstract — During the Covid-19 pandemic, information and the (economic and social) situation has changed rapidly. Traditional (economic) indicators are not sufficiently frequent to monitor and forecast (economic and social) activity at high frequency. We use Google search trends to overcome this data gap and create meaningful indicators. We extract daily search data on keywords reflecting consumers' perception of the economic situation. The indicators are available at www.trendecon.org. An accompaning R package contains the code to construct long daily time series from Google Trends. Robustness of the data is achieved by querying Google mulitple times. The queries are sampled at daily, weekly and monthy frequencies and then harmonized such that the long term trend is preserved. A more detailed methodological description is given on the website. We are currently summarizing these results in a research paper.
COVID_19 Kenya Flexdashboard by Eric Wanjau
Abstract — This COVID-19 in Kenya Dashboard seeks to provide an overview of the 2019 Novel Coronavirus COVID-19 (2019-nCoV) epidemic. The dashboard mainly focuses on Kenya by visualising the total reported cases, the daily reported cases and their distribution per county. It also visualises the total cases globally and then focuses on the cases in Africa and its respective countries. The input data for this dashboard is the coronavirus R package (dev version) by Rami Krispin and the Ministry of Health, Kenya.

Other contributions selected as CovidR contest finalists and presented at the e-Rum2020 pre-conference event are

Nowcasting Italian Covid-19 Epidemic by Marco Mingione, Pierfrancesco Alaimo Di Loro
Abstract — Our work is intended to give people an easy tool for accessing information about the Italian COVID-19 epidemic in an interactive and transparent way. The web application we built is automatically updated every day after the Italian Protezione Civile report (Data Source) and shows both descriptive and modeling analysis, giving the user the possibility to customize several choices.
Pangea - COVID19 in Italy by Nicola Farina, Fabio Priuli, Andrea Romualdi, Martina Tuccinardi et al. - Pangea
Abstract — PangeaCovid19 is an interactive tool that aims to bring insights on the spread of Covid19 in Italy. It is divided in 4 sections in Previsioni we show short term forecasts for the total counts of infected people and deaths obtained with 3 different statistical models; in Terapia Intensiva we show the occupation of intensive care units by covid19 patients on a regional basis; in Per Regione/Per Provincia we show the trend several variables (new cases, deaths, swabs, number of tested people) on a regional basis allowing a comparison between regions; in Matematica della diffusione we explain why we use different forecasting models. Data are provided by Protezione Civile, ISTAT and Ministero della Salute.
COVID-19 Data Visualization Platform by Shubhram Pandey
Abstract — A new invisible enemy, only 30kb in size, has emerged and is on a killing spree around the world: 2019-nCoV, the Novel Coronavirus! This monitor was developed to make the data and key visualizations of COVID-19 trends available to everyone and also provide a platform to conduct a sentiment analysis of social media posts using Natural Language Processing (NLP). This monitor has 3 tabs: Dashboard, Comparison and Sentiments. The dashboard allows the user to view a complete picture of COVID-19 spread around the world. User can also click on any country in the map to view the numbers in that country. In comparison tab, user can compare the spread of COVID-19 in multiple countries in one view. Sentiment tab allows the user to run a sentiment analysis of trending hashtags of coronavirus on social media. The data used in this dashboard is from publicly available reliable and tested source like WHO, John Hopkins, etc.
Understanding COVID-19 through Italian Excess Deaths Analysis by Alberto Danese
Abstract — Why are we expecting to understand COVID-19, when most of the analysis rely on the same biased, unreliable and dull data? Death and cases are counted differently across countries. In Italy, regions and provinces have different capabilities and policies when it comes to testing the population. All this makes drawing conclusions a hard and risky task. Here I suggest an original approach based on overall deaths data from ISTAT, available from 2015 to 2020, done 100% in R. I focus on excess deaths with respect to previous years and answer many questions with a robust quantitative approach. Are men really more hit by COVID-19 than women? Is it true that only the elderly are affected? What really happened in Bergamo and other Italian provinces?
A Coronavirus Tracker, API & Crawler by John Coene
Abstract — A tracker that uses three different sources of data to present the user with a mobile-first application; the app, the API and its data crawler are both open-source.
CoronaDash - Clustering-Analytical Tool for Comparing Countries' COVID-19 Trends by Peter Laurinec
Abstract — The CoronaDash is an interactive dashboard that provides advanced time series data mining tools for comparing countries' trends for any derived statistic(s) with the option of user parameter tuning. For example, comparing countries with “since first” trajectories graphs, clustering these trajectories/multiple statistics with hierarchical clustering, and showing these results with various graphs such as dendrograms, a grid of clusters members, and MDS 2D similarity checks. It also covers and shows various statistics (many derived from the original ones) for every country, and extrapolate cases with exponential smoothing. The data (updated daily) are coming from GitHub repositories of CSSEGISandData, ulklc, and ChrisMichaelPerezSantiago using official data sources.
MINT - Monitoring the Italian National and Regional Contagion Trends by Ilaria Ceppa et al., Kode Solutions
Abstract — MINT is a Shiny dashboard developed to visualize and analyze in a simplified manner data on the spread of the COVID-19 virus in Italy. The web app uses publicly available data, published every day by Protezione Civile Italiana on github. The app automatically updates every day at 6:30pm. In the first tab the dashboard provides an overview of national data, visualizing trends and percent increments of various variables, whereas in the tab “Regional data” the user can select a specific region and explore the data at a regional level. Moreover, in the “Regional Trends” tab, we give the user an overall view of the trends of specific variables across all the regions, while in the “Regional Comparison” tab the user can compare two or more specific regions in terms of their trends and percent increment values over time. Finally, in the tab “Twitter overview”, the app shows the trends for the number of sent tweets and hashtags searches. This data was collected from Twitter from the end of February to the beginning of April.
The 2019-2020 Coronavirus Pandemic - A Timeline by Nico Hahn
Abstract — This app allows the user to scroll through a detailed timeline of key events during the coronavirus pandemic, with charts emphasizing key points in time. It also includes a live map showing the current number of confirmed/recovered/deceased cases for each country. The app uses data from Johns Hopkins University.
COVID-19 Time-Dependent SIR Model by Michele Stefano Altieri, Hypermynds
Abstract — The R package covid19 is used to retrieve, analyse and represent Italian data on the COVID19 outbreak and spreading. It also implements a time-dependent SIR model to predict the epidemic evolution. This package is built from data provided by the Civil Protection department of the Italian Government through the COVID-19 repository.

The following complete the set accepted contributions

100 k African Hospital and Health Facility Locations by Andy South, Anelda van der Walt, Mark Herringer, Ahmadou Dicko
Abstract — An app and R package to improve access to the best-available health facility data for Africa to support the response to the pandemic. There are two main Africa-wide sources of open data on the locations of > 100k hospitals and health facilities. Neither is perfect and there is variable overlap between them. This app allows detailed comparison of the datasets to inform pandemic response and allow improvement. Users can select from 52 countries, view zoomable maps of locations, see summaries by facility type and view the raw data. The R package has functions to make similar comparisons for data obtained from country-specific sources. Part of the afrimapr project, creating R building-blocks to ease use of open health data in Africa.
COVID-19 in Belgium: is it over yet? by Antoine Soetewey
Abstract — This article shows the evolution of hospital admissions and number of new confirmed cases due to COVID-19 in Belgium.
The Prime Minister's Speech(es) by Gabriele Galatolo et al., Kode Solutions
Abstract — Italy was the first European country to face the Covid19 crisis and is one of the most hit by the pandemic. This led the Italian Government to take extraordinary decisions and measures to manage the economic and social effects related to the contagions. In this article we posted on our “Kode Covid19 Blog” we analyzed the speeches given by the Italian Prime Minister Giuseppe Conte comparing them with the reactions we retrieved from Italian Twitter messages having crisis-related hashtags (e.g. #coronavirus, #covid19). With simple text-mining techniques and sentiment analysis we provided visualizations and analysis about the crisis and the population feelings seen from the point of view of both Twitter and the Prime Minister conferences.
OzCoViz Dashboard and the covidrecon Package by Tim Churches (UNSW), Nicholas Tierney (Monash University)
Abstract — OzCoViz is yet another COVID-19 dashboard which nonetheless seeks to go beyond the now ubiquitous world maps and cumulative incidence charts to offer a range of principled epidemiological and statistical analyses, as well as some novel visualisations, including animations. The dashboard leverages the associated covidrecon package which encapsulates much of the data processing required by the dashboard, thus making the dashboard code dramatically simpler and easier to extend. The site has been created by researchers at Australian universities, and hence the focus is on the situation in Australia, within the broader international context, but the site should be easy to adapt for use by other countries.
Europe ETL, API & Tracker by John Coene, Pierre Saouter, Julius Schulte, World Economic Forum Strategic Intelligence
Abstract — While many existing visualizations track the virus’ spread to European countries and provide country-level statistics, the focus of this project is to provide both country-level information for all European states in addition to provincial and regional figures where available. A programmatic pipeline has been built to extract and process the data in a fully automated way, such as to ensure scalability as we add new data sources, full reproducibility, and minimize the risk for manual errors. This also allows us to provide the data back to the community in a structured format through various API endpoints. The overall objective of this project is therefore to present a reproducible framework that could inspire better solutions to the problem of data management for reporting purposes in such crisis situations.
Centralized and Trusted System (QA) to Fight Misinformation by Hicham Moad Safhi
Abstract — Myriad of data is generated every minute from everywhere, and in different formats offering the consumers a large availability of information . However, the reliability of this information differ highly across sources. Unreliable sources provide misleading and biased information, which might lead to dangerous consequences especially in the case of epidemics, like covid-19 in which falsehoods are a matter of life-and death. Users find it difficult to manually check all information, therefore, in this project we propose a question answering tool that can automatically answers the user using a pre-trained model based on reliable data sources.
How to Create a Simple Coronavirus Dashboard Specific to your Country in R by Antoine Soetewey
Abstract — This article will help you to build a visually appealing dashboard about the spread of COVID-19 Coronavirus specific to a country in R using flexdashboard.
Global Covid-19 Explorer by Tom De Smedt, DataHive
Abstract — The Global Covid-19 Explorer is a shiny dashboard using interactive visualizations with innovative metrics to compare the Covid-19 burden in different countries.
coRanavirus early-warning (cRew) Symptom Tracker by Serdar Korur, Dataatomic
Abstract — Important Note >> The app uses Google firebase authentication to work. Thus, only authenticated users are able to interact with the app. (More info and a Demo on youTube.) Covid-19 pandemic caught us off-guard. To avoid this happening again we probably need tools that enable us to react quickly when a new disease emerges and set measures to minimize its spread. In an attempt in this direction, we developed cRew. coRanavirus early-warning (cRew) tracks Covid-19 / flu like disease symptoms in real-time. The goal is to map in real-time healthy and symptomatic people. As users enter data about their health status, the app monitors temporal and spatial changes and estimates sudden increases or decreases on local risks. The app is built with shinyMobile & firebase, echarts4r and fireData packages.
Coro2vid-19 — Leveraging Knowledge About Coronaviruses To Fight COVID-19 by Dennis Hammerschmidt, Cosima Meyer
Abstract — Science-based solutions are crucial – particularly in times of a pandemic. We provide a platform for researchers to find similar articles related to coronaviruses covering more than 22,000 academic abstracts published between 1955 and 2020 from Kaggle. We build our ShinyApp using 100-dimensional word embeddings from Tensorflow through the keras API in R, use Doc2vec from the textTinyR package to estimate similarity scores of abstracts and provide interactive graphs in plotly with clusters of similar articles based on k-nearest neighbors. Users can search for keywords, specific papers, and authors and receive graphs and tables showing related research. The ShinyApp interface is based on the SemanticLibrarian template.
COVID-19 in India - An Analysis by Srinath K R
Abstract — COVID-19 in India - An Analysis is a dashboard/tracker made to understand the spread of the Novel Coronavirus (COVID-19) pandemic across India. The data source is COVID-19 India API, a volunteer-driven API for COVID-19 stats in India which collects data from official press bulletins and reports released by various states. This project provides a summary of all types of cases reported, for the entire country in total and for each state; provides state-wise analysis with the help of a plot and a map and provides both cumulative and daily trends in cases. The main feature of this project is the SIR (Susceptible - Infectious - Recovered) model. This model has been applied for India for a period of 150 days starting from March 1, 2020. Based on this, R0 is also calculated.
Covid'19 Twitter Sentiment Analysis Shiny App by Simileoluwa Kafaru
Abstract — There are so many controvesies and daily trending issues surrounding the Covid'19 Pandemic. Twitter is one of the major Social Media platforms that catches news and trends in real-time, with a wide user base. Here, I focus on a Sentiment Analysis Shiny App, capable of collecting tweets in real-time with a choice of major hashtags. Sentiment Analysis is performed on the retrieved tweets and dashboarded. Also, a daily stats of the Covid'19 pandemic is presented, all with real-time updates.
Dashboard to Monitor Covid-19 Pandemy in Italy by Vito Fanelli
Abstract — This dashboard allows to monitor Covid-19 pandemy in Italy. It is based on daily data of Protezione Civile. The “Italy” section shows the number of total cases, actual cases, healed and deads at country level. It contains daily trends, moving averages and mortality rate. In the “Region” section there are the same information at regional level.
COVID-19 Canada Data Explorer by Petr Baranovskiy
Abstract — COVID-19 Canada Data Explorer is an interactive Shiny app to break down and analyze the official COVID-19 dataset available from the Government of Canada (https://health-infobase.canada.ca/src/data/covidLive/covid19.csv). The app can be used to visualize 13 different indicators by time and geography, and to directly compare the epidemiological situation across Canada's provinces and territories. It also allows users to custom-build and download their own datasets.
Scientific Papers Related to the COVID-19 Virus by Juan Sokil, Observatorio-CTS
Abstract — This Shiny web app provides information updated daily on scientific papers related to the COVID-19 virus. The data is extracted from PubMed, a free search engine accessing primarily the MEDLINE database of references and abstracts on life sciences and medical topics. This app allows to analyze the scientific production by country, the scientific collaboration and the topics of research.
COVID-19 and R Blog Series by Tim Churches
Abstract — A series of blog posts starting on 18th February 2020 demonstrating the use of R to obtain, analyse and visualise COVID-19 data, including scraping of detailed data from wikipedia, the calculation of the time-variant effective reproduction number \(R_{t}\), which has subquentially become widely adopted as a measure of COVID-19 intervention success, and the creation of both dynamic and stochastic individual contact models to simulate COVID-19 spread and investigate the likely effects of various public health interventions in a population.
COVID-19 Global Cases - Open Source Dashboard by Christoph Schoenenberger
Abstract — This dashboard shows recent developments of the COVID-19 pandemic. The latest data on the COVID-19 spread are regularly downloaded from Johns Hopkins University and displayed in a map, summary tables, key figures and plots.
Real-time Interactive Global Distribution of 2019-nCov: Cases, Deaths, and Cases in Last 15 Days by Matt Malishev
Abstract — Interactive map of coronavirus 2019-nCov global distribution using live webscraped data from the European Centre for Disease Prevention and Control (ECDC).
Data are for the country level showing cases, deaths, and cases in last 15 days, and show individual country rankings for these metrics.
Covid19 Data and Perception Application (CoDaP) by Sabina Rako
Abstract — Covid19 Data and Perception (CoDaP) is a Shiny web app created based on open results of International Survey on Coronavirus (https://covid19-survey.org/) and open epidemilogical data from European Centre for Disease Prevention and Control (https://www.ecdc.europa.eu/en).
Covid19 Shiny App by Francesca Vitalini, Mirai Solutions
Abstract — Covid19 is a Shiny web app designed for visualizing trends, comparing the situation in different countries and getting a global overview of the virus distribution. The application uses data from the Johns Hopkins CSSE Repository untill March 23rd and Worldometer data afterwards.
CovidR Contribution Example by An Author, Another Author
Abstract — Short abstract with a brief description of the contribution and its main features (max 800 characters, including spaces). The usage of publicly available data should be explictly mentioned. The text can be broken across multiple lines, each indented with two spaces.