Since its inception in 2009, a main goal of the RStudio community has been improving the accessibility of R within and beyond data science. Education is an important component of that plan. To help extend the capabilities of their internal teams, RStudio has helped create a network of partners and teachers to help develop and implement guided data science training and workshops.
Last year, RStudio launched a formal certification process for instructors that are developing materials and teaching data science with R and RStudio. Prospective students, corporate employees and individual learners alike, can select from a certified group of trainers that have been vetted by RStudio.
At a high-level, the certification process includes three steps:
- Training course on modern teaching methods
- Exam and demonstration on education principles
- Exam and demonstration on technical competencies
To find out more information on the new certification program, we spoke with Greg Wilson, who leads the education group at RStudio.
Launching the Certification Program
In the 1980s and early 1990s, Greg was a programmer at a physics department, where he noticed a recurring pattern of scientists spending an unnecessarily large amount of time on problems that could be quickly solved programmatically. This common challenge led Greg to co-found Software Carpentry, a volunteer non-profit organization designed to teach basic coding and data science skills to researchers and scientists.
When Greg started teaching people how to teach, he realized there were big gaps in his own understanding, so he went back and started filling them in with evidence-based research on education theory. Before teaching the basics of programming, it was important that instructors understood how people learn and how to best design a workshop for knowledge retention.
Greg joined RStudio in September 2018. Shortly after joining, he was tasked with focusing on the RStudio education program. In that role, he applies the same Software Carpentry principles and methodologies to RStudio instructor training.
RStudio recognizes the limitations in trying to educate and train the community with only internal staff, so the overarching goal of the education program is to help community-based and partner trainers be more effective. These certified trainers are then enabled to incorporate their own knowledge, experiences, and subject matter expertise, while leveraging smarter teaching methods, to develop RStudio training programs.
Evolution of the Certification Process
The new RStudio certification program launched in January 2019 at rstudio::conf in Austin, TX, with the first certifications offered for Shiny and the Tidyverse. The first certified instructors started graduating the program in April 2019. Since that time, almost 100 people have been certified.
To better support the growing community of R and RStudio users, the training courses are taught via interactive online learning. This also allows the instructors, who are typically familiar with the traditional in-person learning format, to experience virtual learning best practices.
Modern Teaching Methods
One core component of RStudio’s certification program is the integration of modern, research-based teaching methods into the creation of data science curriculum and workshops. There’s an extensive body of research that covers these methods in great detail, but at a very high level, this involves adapting course materials for individual circumstances and encouraging peer instruction.
Studied and popularized by Eric Mazur at Harvard in the 1990s, peer instruction is the concept of having students argue their viewpoints with each other through interactive learning. This interaction has been proven to increase overall knowledge retention because it scales individual intervention in a way that teacher-student interaction can’t support. While the peer may not be able to explain as well as the teacher, they can still explain individually.
Another modern teaching method used throughout the program is formative assessment. This practice involves check-ins after the introduction of every major concept or technique (typically every 10-20 minutes). If a learner doesn’t fully grasp a concept, the teacher can correct misunderstandings right away rather than going on and leaving people behind.
“In the long run, we’re looking for more people to know what good education looks like and why. And this spans all forms of education, not just data science.” -Greg
For more information on peer instruction and an overall collection of Greg’s learnings on effective teaching methodologies visit https://teachtogether.tech/. Many of these concepts are incorporated into the certification program at RStudio, designed to guide certified instructors in the development of their own programs.
The R community is excited about the new program and the reception has been very positive. With the interactive nature of the training program and the small class sizes, there are close to 400 people on the certification program wait list.
Geographic coverage for instructors is currently heavy in North America and Western Europe. RStudio is looking to expand that footprint to better serve instructors and learners on a truly global scale. Expanded language offerings for teaching courses and exams are being developed – currently for Spanish and French.
As previously mentioned, the program currently covers the Tidyverse and Shiny. RStudio is constantly looking for ways to improve and expand this program, particularly in light of the mass movement of teaching online on short notice. If you have thoughts on new offerings that you’d like from the RStudio education team, let us know in the comments below or reach out to us and we’re happy to share with RStudio.•