CRAN: the Granddaddy of Analytical Marketplaces
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
by Bill Jacobs, VP Product Marketing, Revolution Analytics
I had a most interesting exchange with an industry analysis firm recently who suggested that application marketplaces were critical to the success of analytical tools, suggesting that Revolution Analytics was remiss in not creating one.
I must say, I was taken aback somewhat. In considering the suggestion, I was left suspecting that the expressed enthusiasm for marketplaces being built to do for various commercial products what the R community already enjoys.
We responded that the R community already enjoys a richly-furnished “marketplace” for R extensions, algorithms, applications, adapters, techniques and educational assets. I’m speaking, of course, of CRAN and the 6000+ CRAN packages.
Is CRAN a marketplace? Perhaps the typical goals of a marketplace hold an answer:
- Foster success among groups and individuals working to enhance the capability of a particular tool or solution;
- Provide a “first place to look” for users tackling new problems;
- Foster business, technique and industry-specific communities of practice;
- Distribute pre-built solutions for specific use cases;
- Provide a vehicle for third-party contributors to monetize their efforts if desired.
Consider CRAN in light of these goals:
- CRAN supports the entirety of the R community and is in near universal use among R users.
- CRAN succeeds to a level unmatched by any competing marketplaces. The sheer number of innovative packages contributed to date far outweighs anything of its kind elsewhere in the analytics community.
- CRAN provides vast numbers of pre-built application components including freely shared algorithms, integrations, transformations, visualizations, data connectors, educational assets, and the list goes on.
- CRAN hosts innovation in algorithms, easily visible by watching authors continuously improve packages in response to community inputs.
- CRAN enables collaboration on vertically-specific problems — one need only search on a single field such as genomics to see examples of communities of practice sharing their work via CRAN.
- CRAN is big and growing vigorously. A couple of week ago, I surveyed new contributions to CRAN in first 2 ½ weeks of 2015. The list included hundreds of new packages for genomics and life sciences, chemistry, natural resources, biology, ecology, forestry, agronomy astronomy, drug research, healthcare delivery, finance, forestry and government.
If R users struggle with CRAN anywhere, they do so perhaps as a victim of CRAN’s huge success. With 6285 contributed packages as of this writing, CRAN can be unwieldy for the new users. And a solution exists. New users often rely on the wisdom of the R community, as expressed through the “base” and “recommended” packages that are typically installed with R. These provide a rich, but tractable “starter set” of frequently used techniques, algorithms, connectors and other methods.
Among the typical goals of a marketplace is one not targeted by CRAN. CRAN is not used to monetize directly, reflecting its open source heritage. CRAN provides free distribution to end users, whom I suspect, generally find “free” to be a great advantage.
CRAN can no doubt be improved and extended. For example, CRAN mirrors the latest versions of most packages, introducing potential repeatability problems when sharing scripts. We’ve built what we call the Managed R Archive Network (MRAN). MRAN extends and takes daily snapshots of CRAN, addressing issues of reproducibility, while continuing to expose the entirety of CRAN.
And we eat our own dog food. We use MRAN to distribute Revolution R Open, our new and accelerated version of open source R, with source code. We also distribute introductory material on R targeted to new users and provide significantly improved searching tools.
MRAN lives at http://mran.revolutionanalytics.com. Have a look.
At Revolution Analytics, we do not believe the R community needs another marketplace for packages. We believe that CRAN amply fills the role, perhaps better than any vendor-specific application marketplace available today. With extensions such as MRAN and the repeatability toolkit, it excels at its job.
One need only compare CRAN to any other marketplace. Look for the number of packages contributed independently. CRAN is a shining example of the open source business model operating as intended. CRAN continues to foster a thriving community with shared interests that is building innovative additions to the R ecosystem including components and solutions, but also educational assets, data assets, and many others.
Sounds like a pretty effective “marketplace” to me…
R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.