I last got an update on RapidMiner back in March 2012. The company was founded in 2007 and has 40 employees worldwide. The free product has reached 10,000 downloads a month and the product has some 35,000 active deployments with over 500 customers worldwide (40 added in the last 6 months). The company recently secured a $5M series A funding round and changed its name to RapidMiner (to unify the product and company) They have also launched a RapidMiner marketplace offering analytic extensions to the product. These vary from open source additions to more commercial ones and the size of their user base is a critical driver for the range of solutions in this. Just over two thirds of their customers are from EMEA with about a quarter coming from North America.
RapidMiner focuses on supporting predictive analytics with “no programming required” and has been eliminating the need to write code or script in the analytic workflows. RapidMiner is one of the highly ranked vendors in the Rexer Analytics survey – one of the top 3 most used tools. They feel that the drivers for this are:
- The open source community and marketplace
- The integration of the platform and its focus on not requiring programming
- The scale of the platform in terms of connectivity and algorithms
- Its support for strong version control
- Its one click deployment of any analytic process as a web service for integration into deployment environments
RapidMiner 6 was released a few weeks ago and has a number of new features focused especially on helping new users get rapidly to a predictive analytic solution. In particular they have added support for prebuilt templates or wizards for self-service. These allow a non technical user to select their industry, focus area etc and get a list of template approaches. They can then plug their data into the wizard and get a first cut process to address the selected problem. This will provide both the initial model, some reporting on performance and impact as well as a “story” that describes the analytics in a way that makes sense to a non technical consumer. 4 shipped initially and RapidMiner intends to add to these while also supporting third parties developing these and making them available on the marketplace.
In addition they added
- New welcome screen to get started faster
- New result view with call to action
- New statistics and charts views
- Improved tours and tutorials
The licensing model for v6 also changed, simplifying it significantly:
- Free Starter with a 1Gb data limit and access only to files
- Personal for $999 with some support a 4Gb limit and access to open source databases
- Professional for $2999 with some support, 8Gb limit and access to all databases
- Enterprise with full support and no limits
One of the key features of this new license model is that all the analytical features are available at every level.
RapidMiner studio is still a fat-client, java-based environment focused on visual layout of analytic workflows. It continues to offer strong checking of the flows to identify potential problems (based on analysis of the flow and its meta data) as well as lots of operators that can be manipulated entirely graphically.
RapidMiner Server supports large scale deployment of the analytical processes as well as conversion to web services. Various dashboard and visualization elements can be displayed also and every element of the server dashboard is based on a RapidMiner flow. This allows RapidMiner processes to be used to control the behavior of the server in terms of invoking other services, sending messages or updating databases for instance. This would allow you to develop a full range of model management capabilities to schedule and monitor model updates etc.
From a roadmap perspective they are focused on improving access to cloud data sources and cloud-based execution of processes and delivering results more effectively to mobile devices.
RapidMiner is one of the vendors listed in our Decision Management Systems Platform Technologies report and you can get more information on the product at rapidminer.com.
Comments on this entry are closed.