We have stressed the importance of big data visualisations a few times on this blog. Ayasdi, which is a spin-off of a DARPA-funded Stanford research project and maintains close ties to the Stanford math department, created a platform called Iris to visualize entire massive data sets, rather than smaller queries and slices. They use a mathematical technique called topological data analysis to find unexpected insights for the pharmaceutical industry, the energy industry, and others.
Ayasdi describe the product as a cloud-based machine learning platform that uses visualizations of massive data sets to discover unexpected patterns and connections. They already made some interesting statements in a earlier stage based on their methodology. At the MIT Sloan Sports Analytics Conference, Ayasdi analyst Muthu Algappan argued that there are really 13 positions in basketball, instead of five. Ayasdi’s software claims that basketball coaches are better off looking at their players as “shooting ball handlers” and “scoring rebounders”, instead of centers and point guards.
Another application of Ayasdi comes from the Netherlands. Using 11 years of data mined from the Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, Ayasdi was able to identify new, previously undiscovered populations of breast cancer survivors. Using connections and visualizations generated from the breast cancer study, oncologists can map their own patients’ data onto the existing data set to custom-tailor triage plans. In a separate study, Ayasdi helped discover previously unknown biomarkers for leukemia.
“The biggest challenge in big data today is asking the right questions of data. The power of Ayasdi is its unique ability to automatically discover insights without asking questions”, said Ayasdi CEO Gurjeet Singh. This is a very interesting statement regarding big data analysis. Big data visualizations might be the most important trend to bring big data to the business side of things. Non-data scientists interacting with data without having to ask questions might be something that will speed up big data practices in 2013.