-
-
cumulative cases by federal state
-
daily new cases by federal state
Inspiration
There are a number of R packages and Web-Apps about global COVID-19 epidemiological data, but so far, to our knowledge, no package that provides an easy and ready-to-analyse access to the German Robert Koch Institute (RKI) dataset. In particular, most other global datasets lack resolution on the German "Landkreise" (Counties) and "Bundesländer" (federal states).
Our goal was therefore to provide access to the german epidemiological dataset by the RKI and showcase this access with example visualisations, both static and interactive in form of a Web-app.
What it does
Features
- Loading the most updated dataset from the RKI.
- Caching of data for faster loading and to enable offline-access
- Easy grouping of data by Country, Federal State and County
- Direct Access to key meta-data such as number of inhabitants per county and number of hospital beds, to enable creative data analyses.
- Simple visualizations of the data, both static and interactive
- Simple extrapolation of trends into the next 7 days
Package
We provide a core package, hosted as an R package on GitHub. The core package provides key functions to load, group and visualise the epidemiological data from the Robert Koch Institute. There is an automatic caching function to enable working offline and to efficiently load the data on a daily basis. Data is readily represented as a tibble, i.e. a data frame with columns for positive test cases and deaths per county. In order to visualise, we opted to provide the user with code snippets and examples that showcase the packages, rather than provide dedicated functions for individual plots. We cannot envision what other users might want to visualise, so we rather enable users to start playing around with this data themselves.
Web-App
To showcase interactive usage and visualisation of the data, we provide a prototype for a Web-App, based on R's shiny framework. The app code is hosted in the package repository and the app itself is available here.
The shiny app showcases an interactive example how to use the data, and provides an easy-to-access way for people outside of the programming community to visualise the RKI data.
How we built it
Core technologies of the package are the packages in R's [tidyverse], one of the most widely used and well-documented packages for data analysis in R. The Web app is based on Shiny, which features a huge open source community.
Challenges we ran into
Things went rather smooth, as we focused on creating something we hoped to be well usable after a limited time frame of 48 hours. We were right.
Accomplishments that we're proud of
We can provide a useful tool for a wide range of users with interest in the development of the Covid-19 pandemic, from interested citizens to data analysts and decision makers. Providing a package for the popular R statistical language, amended by a web app, we achieved a very low barrier to entry.
What we learned
We all learned a lot from the other team members.
What's next for 0999_ZahlenundKurven_Prognosedashboard
A lot of features for the web app like a map interface and various statistics like doubling time or infection or different forecast methods are possible.
The R package itself will be extended by some further data sources and features, but we think it should keep the clearly defined purpose of retieving and preparing data related to the current Covid-19 pandemic. Data analysis is highly diverse and should live in other packages.
Log in or sign up for Devpost to join the conversation.