Virtual useR! 2024

Chemotion-ELN is an open source electronic lab notebook (ELN) developed by the the ComPlat group at KIT in collaboration with the consortium NFDI4Chem within the National Research Data Infrastructure. This session delves into the seamless integration of Shiny apps with Chemotion-ELN, enabling users to effortlessly transmit data to these apps for analysis. The results are then seamlessly conveyed back to the electronic lab journal, enhancing its capabilities. Through this bidirectional communication, the ELN becomes highly extensible, empowering researchers with a versatile platform. In this presentation, we spotlight Biostats, an exemplary Shiny app. This application is tailored for fundamental statistical tasks, encompassing data wrangling, visualization, and the multifaceted nature of statistical tests. Witness the transformative potential of connecting Chemotion-ELN with Shiny apps, exemplified through the intuitive functionality of the Biostats app.

The success of the R programming language is largely due to its ease of creating and sharing R packages. We propose an opinionated framework called “openstatsguide”, published on openstatsware.org/guide.html, which can guide R package developers towards a minimum set of good practices. As far as we know from our literature search, this is the first attempt at providing a small and concise set of rules for package developers. This applies not just to R, but can also be used for functionally oriented programming languages used in data science, and we give examples for R, Python, and Julia. Rather than a full and detailed how-to guide, we keep “openstatsguide” short and on a high level, thus lowering the entry point for novice and seasoned developers alike. Our hope is that this guide can increase the adoption of software engineering good practices in the statistics community. In this talk we describe the motivation and scope of “openstatsguide”, relationship with existing work, the set of good practices, the maintenance model and ideas for future complementary guides produced by the openstatsware.org working group.

The buzz around the {shiny} package has only increased over the past year. Now, more and more R programmers use interactive dashboards to share their data insights with a diverse stakeholder audience – from their colleagues to project managers and sponsors. But even for seasoned R programmers, any dashboard or application development is a resource intensive process and, if not planned carefully, can cause the team delays on the project deliveries, or worse, suboptimal user acceptance rates. Quick and targeted prototyping can prevent many of these situations from happening. This talk is aimed at fellow data scientists who want to start implementing {shiny} applications in their workflow but have little to no background in web application development. In this presentation, the author shares lessons learned from their own mistakes during the development of various production ready {shiny} applications to help novice developers avoid similar pitfalls. The attendees will see strategies for creating and testing a quick user-interface mock-up design using {fakir} and {shinipsum} packages. Additionally, they will learn how to translate the prototype into a minimum viable product design.

Social and ecological systems around the world are experiencing multiple crises [1]. Decision makers have shown a lack of ability or will to take action to reduce current social and ecological injustices [2]. However, individuals without formal political or economic power can contribute to peacefully guiding the world back to a sustainable pathway. The Dandelion Hub (https://dhub.global) serves as a central repository for de-central, non-violent actions for ecosocial justice. The Dandelion Hub (DHub) uses a web frontend (R/Shiny) and database backend (MariaDB) to record and report actions across a broad spectrum of non-violent actions according to the classification by Sharp (1973) [3]. As of 2024-01-28, 2,389,238 activists across 615 actions in 180 cities, 50 countries, and 6 continents, who took part in non-violent actions between 2018-08-20 to 2024-01-14, are represented in the repository. Open-source technology, such as R/Shiny, can contribute to the socio-ecological transformation of society. References: [1] WEF (2024). http://tinyurl.com/2995u9jh [2] IPCC (2023). http://tinyurl.com/5cybutfs [3] Sharp, Gene (1973). http://tinyurl.com/2jvhvxwt

Cohort analysis is vital for understanding patterns and trends within datasets, particularly in fields like healthcare, marketing, and user analytics. The 'cohortBuilder' and 'shinyCohortBuilder' packages in R offer a convenient approach to defining and manipulating cohorts. If you're exploring ways to streamline your cohort analysis workflow within Shiny, this talk will introduce you to powerful tools worth exploring. During the presentation, I will demonstrate the core concepts of the 'cohortBuilder' ecosystem, highlighting their strengths in performing cohort analysis and visualization within R Shiny.

FODESAF is the main financial instrument of selective social policy in Costa Rica, in the fight against poverty. From 2005 up to 2022, according with the evolution of FODESAF´s cash flow, its income was around US$ 13 thousand million, resources that have been transferred, nearly in a 100% basis, to social programs nationwide. Only in the year 2022, the total amount of money transferred to the market was around US$ 1.000 million. Such massive amount of resources of FODESAF´s cash flow has been analyzed due to the design of a Data Pipeline in R. Its inputs stage (ingest) are all reports and data coming from technical-internal offices and external institutions. The processing stage is constituded by data mdeling and tidy packages in R. The final outputs (results) are HTML reports in Quarto© for managerial levels and government officials, as well as Power Bi© reports inserted in FODESAF´s webpage for general consultation of citizenship as a whole, in order to achieve transparency mandates. Due to the usage of R, large amounts of data are processed to generate high-quality final products, in order to support the design of better public-social policies.

With the release of webR, the barrier of entry to use R on the web has been significantly lowered. Through embedding webR within Quarto, a novel form of report interactivity has emerged, giving rise to pseudo-web applications. These applications span the spectrum from concise tutorials to intensive edge computing scenarios. In this talk, we'll discuss the transformative paradigm shift within interactive data science facilitated by the Quarto extension {quarto-webr}.