Early on in my career at SGI we built some interesting data management tools and data warehouses for processing world wide business data and BILL INMON was my hero(I am sure he was the hero for a lot of data geeks back then!).
This recent article on BIG DATA: SEARCH VS. ANALYSIS by him, is a great validation of the need for structure for analysis.
"All data in big data is unstructured. Why is this important? It is important because with unstructured data you have no metadata. You have no context with unstructured data. This means you can only do the simplest of searches against unstructured data. Since there is no context, you can’t do any real analytical processing." – Bill Inmon.
At Glassbeam, this has been our core philosophy. For machine data (such as produced by devices) around the internet of things, one needs to create some structure, even if its dynamic, before embarking on analysis. Our core platform and language is designed from the ground up for business intelligence on machine log data which is primarily unstructured.
With Glassbeam our customers get the best of both worlds, in being able to search on unstructured data as well as analyze structured data with Glassbeam Analytics. Glassbeam can do this because it applies parsing rules, metadata and structure before loading unstructured data.
SEE MORE ARTICLES ON SEARCH VS. ANALYTICS on log data or unstructured data by our team on this topic.