Rich.Pugh <- presidency.r.consortium

Congratulations to one of our own ISoP members, Richard Pugh of Mango Solutions, who has been selected to be the Inaugral President of the R Consortium.

The R Consortium, Inc. is a group organized under an open source governance and foundation model to provide support to the R community, the R Foundation and groups and individuals, using, maintaining and distributing R software.

As per the blog post released on the R Consortium website Richard will lead the newly elected Board of Directors consisting of members from a number of well known organizations :

  • JJ Allaire, RStudio
  • Louis Bajuk-Yorgan, Tibco
  • John Chambers, R Foundation
  • David Smith, Microsoft

Our indefatigable ISOP reporter caught up with Rich recently for this interview.

Reporter: Congratulations on your role as the new president of the R consortium: can you tell us more about the R consortium and how it may help individuals in the Pharmacoemtrics community?

Thank you. The R Consortium was formed in 2015 by a group of companies who are fans of the R language, as a non-profit Linux Foundation initiative. Mango have been huge advocates of R for many years, and have often delivered R training courses at the ACOP conferences. A key aim of the consortium is to support organisations as they instill R as a primary analytic technology. R is increasingly popular within the Pharmacometrics community, so I hope the Consortium will be able to support the future adoption and development of this fantastic technology.

Reporter: As the first president of this consortium, what do you hope to achieve in your term?

I’m incredibly proud to have been elected as the first President. The President’s role rotates between the members of the R Consortium Board, with each member serving for 6 months. As such, my first term is almost up and I’ll shortly be handing over the role to Lou Bajuk-Yorgan of TIBCO. As the first President, my role has involved a lot of “startup” activities, setting up the Infrastructure Steering Committee (the mechanism by which the Consortium awards grants to projects that will benefit the R Community), and discussing the benefits and challenges of using the R language with a range of organisations.

Reporter: As president, what learnings form the PMX world will you bring to the R consortium?

I’ve been a member of the Pharmacometrics community since 2001 when I attended the PAGE conference in Basel. I was immediately struck by the dynamism of the industry, especially when compared to my background as a Statistician using a very popular, yet less flexible (!) analytics tool. The thing that struck me then, and I still enjoy, is the ongoing drive to continue to innovate to push the industry forward. In that way, I think PMX and R are a match made in heaven and it’s no surprise to me that R is so popular in this community. However, the PMX world has taught me lots about the challenges of using open-source, “flexible” tools in a heavily regulated industry. This is certainly something that is at the heart of everything I’ve worked on in my role as the R Consortium President, and will continue to heavily influence my actions in the industry.

Reporter: R has played a large role in our community for data manipulation and visualization for a number of years. We are also now seeing more and more R packages that support various analysis aspects in the Pharmacometrics community, such as the PMXstan package, RxODE, nlmixR, etc. As President, where do you see R in the future in relation to other analysis tools currently used in the community?

Right now, I’m seeing a massive growth of R as a key tool for analysis & visualization. I think this trend will continue, but I also see other possible technologies that could add significant value to the PMX arena, such as Python and Julia. However, I think a large part of the “future” of modelling in general (both in PMX and elsewhere) is how R has managed to evolve thinking about what programming means within the analytic sphere. I think R pushes us to think more about what “good programming” looks like, enabling us to switch to new technologies more readily in the future as they become available. As for the “big data” technologies such as Hadoop and Spark, I don’t yet see a need, but this may change as larger and more varied data sources become available to us.

Reporter: As president, how do you see either 1) ISOP of 2) individual pharmacometricans helping to play a role in supporting the R consortium?

I see the R Consortium’s role is primarily to support the R Community, so I would urge ISOP and pharmacometricians to challenge us to drive the evolution of R to make it an even better tool for modellers. I’m happy to discuss how you’d like to see the language develop so please do drop me a line (rich@mango-solutions.com)!

Reporter: In your role as president do you see anything coming in the R community that may have an impact on our discipline?

I think I see many changes in the R language, both over recent times and “coming soon!”. For example:

• We seem to be using R now at a “higher level” (if that’s how to describe it). When I started using R (and before that, S+) my life was spent grappling with dollar signs and square brackets. With the advent of packages like dplyr and magrittr, we seem to be moving (from a syntax perspective) out of the basement and up to the first floor somehow. This makes the language elegant and more readable, making it more accessible to less-technical colleagues. I’m already seeing the implications of this, delivering training courses to users who would otherwise be using Excel, using a coding style that is far more accessible.

• I think the Shiny platform from RStudio has been a game-changer, and I continue to be impressed with the simplicity with which we can deliver interactive results to business stakeholders

• Similarly, the advent of RMarkdown seems to have made markdown-based reporting far more accessible, and we spend a lot of time helping users to create reproducible reports and outputs

• The forthcoming R-Hub project, and similar initiatives, point to future where R is more readily managed and the “quality” and “usefulness” of packages is scrutinised even more. This can only help to drive standardisations across package usage and ultimately provide us with better code.

On behalf of ISOP, congratulations Rich!