“11Ants Analytics is determined to make data mining technology accessible to less-technical users. Ordinary people like me and you. They got rid of the intimidation-factor most people like us experience when attempting data mining by integrating their Model Builder into Microsoft Excel so it also provides some seamless workflow conditions.
Their 11Ants Model Builder gives users access to a range of advanced analytical tools and features, in an easy to use navigation that allows less-technical users to quickly extract patterns from data and build predictive models. The software emulates the best practices of a data mining expert and does the complex work for users, effectively a data mining auto-pilot.
The company began its business narrative as Khipu Systems in 2005. It was a joint venture between Hamilton-based Hill Laboratories and Waikato Link (the commercialising arm of Waikato University). The company leveraged data mining and predictive technologies to deliver quicker and more accurate soil testing for quality control and compliance purposes.
When Hill Laboratories purchased the rights to that technology, Khipu began building other niche applications to help decision makers cull informational gems from the mega-data downloads of scientific instruments. The prototype software was inspired by machine learning breakthroughs at Waikato University and was built using the Waikato Environment for Knowledge Analysis (WEKA) open source toolkit developed by the University’s Computer Science Department.
As a bigger picture emerged, the company leveraged its own developments to create a ‘data agnostic’ approach. The main goal of this approach is to provide an environment within the company with the capability of accessing data to meet their needs in an easy manner. That meant that the ‘look and feel’ provided by the technology defined would be very similar throughout organisations if not exactly the same. The new owners changed its name to 11Ants Analytics to better reflect this new direction.
“People will try to avoid learning something new unless they have a burning need. A11 Analytics wanted something they could pick up and learn quickly,” said Tom Fuyala, Director, Business Development for the company.
It was clear most projects started in Excel and after the data was exported for analysis ended up back in Excel, a package most respondents were comfortable with. That drew a compelling picture. After determining whether it was possible to make very complex back-end processes accessible through Excel, work began on ways to achieve this without forcing people to learn lots of new commands.
THE SECRET SAUCE
The ‘secret sauce’ at the heart of the 11Ants Model Builder Excel add-in is HyperLearn technology, which emulates how a data mining expert might use a range of complex algorithms. Over three years the 11Ants team compiled, wrote and assembled a library of 11 high-level public domain algorithms, typically used in academia or the IT departments of major global banks.
They’ve wrapped them up in HyperLearn so people don’t have to learn all these skills. Even the experts like it because it can be operated under supervision and saves a lot of time. Their goal was to bring modelling closer to less-technical people who already understood the data they have so they can conduct analysis without bringing in a specialist programmer.”