What Does R Integration Really Mean For BI Platforms?

I just received yet another call from a reporter asking me to comment on yet another BI vendor announcing R integration. All leading BI vendors are embedding/integrating with R these days, so I was not sure what was really new in the announcement. I guess the real question is the level of integration. For example:

  • Since R is a scripting language, does a BI vendor provide point-and-click GUI to generate R code?
  • Can R routines leverage and take advantage of all of the BI metadata (data structures, definitions, etc.) without having to redefine it again just for R?
  • How easily can the output from R calculations (scores, rankings) be embedded in the BI reports and dashboards? Do the new scores just become automagically available for BI reports, or does somebody need to add them to BI data stores and metadata?
  • Can the BI vendor import/export R models based on PMML?
  • Is it a general R integration, or are there prebuilt vertical (industry specific) or domain (finance, HR, supply chain, risk, etc) metrics as part of a solution?
  • What server are R models executed in? Reporting server? Database server? Their own server?
  • Then there's the whole business of model design, management, and execution, which is usually the realm of advanced analytics platforms. How much of these capabilities does the BI vendor provide?

Did I get that right? Any other features/capabilities that really distinguish one BI/R integration from another? Really interested in hearing your comments.


I really like this change

As someone who's certainly not a BI expert, this is the first I've really read about this. It excites me because think R is one of the best open-source packages of any kind, and it's great to see it push further from academia to enterprise.

Use case integration

I think another potentially distinguishing feature of R integration (or predictive analytics integration in general) is how well it is integrated in existing LOB/use case specific packages offered by the BI vendor, such as financial planning. This greatly reduces the skill set required by the customer, and reduces time to value through the use of "pre-canned" data models and a pre-selected set of algorithms

Good point, thanks, adding to

Good point, thanks, adding to the list

Integrating R in general and into Alteryx in paricular


Leading the team that integrating R into Alteryx, I have had a chance to talk to a lot of our customers, and they always raise many of the same questions around R integration and functionality that you raised in your blog. What we are seeing is that there are some key capabilities that are needed before the use of R inside of mainstream BI/Analytics platforms crosses the chasm to general use across a wide range of use-cases:

(1) Ease-of-use – though in my opinion it’s not just about point-and-click but also involved some guidance that helps analysts who really understand their data and business to start ramping up their use of R without hitting a lot of dead ends or making basic mistakes. This should ideally include the ability to go through the analytic development model (design-management-execute-iterate) from within the BI/Analytic platform.

(2) Integration, integration, integration – couldn’t agree more that R capabilities executed in isolation can be sub-optimal. The best approaches will be those that integrate R into the BI/Analytics platform bi-directionally.

(3) Leverage existing expertise/investments – this should mean the ability to leverage existing models (with the ability to import them into the BI/Analytics platform relatively intact). It could also mean the ability to leverage existing investments in servers/infrastructure – without having to procure/provision a new server to execute R models.

I've written a blog post that directly dicusses how Alteryx is addressing the four issues of R integration you raised. The post can be found at: http://www.alteryx.com/community/blogs/engine-works/ask-dr-dan

R is a step backwards for BI

Let's remember that R is a programming language. It is cool because it is a very comprehensive programming language designed for statisticians by statisticians. But, it is not designed for business users to be productive. Using R for BI would be like using Java language for BI, or maybe even RPG (report programming language). It has it's place, just as programming languages has its place. But, the future of analytics should be easy to use visual tools for business people.

Leads to the "What is BI" flop-and-twitch discussion

See your point, Mike, but I tend to regard data mining and predictive analytics as part of a broader BI function, even if it still in the realm of statisticians and data scientists. There are a number of vendors working on making PA-for-the-rest-of-us type solutions that will eventually make the wizard's practice more commodity.

There are vendors that make data mining...

...much easier than coding R now. The Predictive Analytics Wave that Forrester will publish in about one month evaluates some of those PA-for-the-rest-of-us solutions. I agree with you that data mining has a big place in BI.

Looking forward to the wave

I've had a number of inquries that I wanted to shuffle your way because they wanted the wave data.

Coding BI

Agreed. Generally I find that, in any visual development environment where people have the choice of a point/click solution versus a coding solution, they tend to wind up with too much code! Much of the time they have reinvented the wheel. The best solutions seem to be the ones where the code functionality is only intended to fill in gaps in the visual design capabilities.

Goes back to Boris's first point...

"Since R is a scripting language, does a BI vendor provide point-and-click GUI to generate R code?" R doesn't have to be a step back if BI & Analytics vendors treated it less like a check-box, Mike. The Alteryx team ended up taking the most commonly used statistical & algorithmic functions & created good UI around it, so you won't have to program in R to taking advantage of it. Totally agree that it's all about visual tools!

R Integration

Q: "Did I get it right?"
A: Not quite... There are some easier ways of doing it, i.e. HTTP, XML, JSON, or even CSV. All these can be done it run-time mode