Therefore, he needs tools that allow him to freely explore and investigate data. Incorporating more data sets in the analysis should be very easy, there should be no need to specify a goal beforehand, and it must be possible to analyze data in an unguided way.
Besides all the more standard features such as displaying data as bar charts, in dashboards, on geographical maps, the perfect tool for this type of work should support at least the following characteristics:
- No advance preparations: There should be no need to predefine data structures of the analysis work in advance. If data is available, it should be possible to load it without any preparations, even if it concerns a new type of data.
- Unguided analysis: Analysts should be able to invoke the analysis technology without having to specify a goal in advance. The exploration technology should allow for analyzing data in an unguided style.
- Self-service: Analysts should be able to use the analysis techniques without help from IT experts.
CXAIR internally organizes all the data using an intelligent index. In fact, internally it's based on text-search technology. This makes it possible to combine and relate data without any form of restriction, which is what analysts need.
Unlike most analysis tools, CXAIR uses search technology that speeds up data analysis. For calculations a mixture of in-memory and on-disk caching is used to analyse massive amounts of data at search engine speeds. All the loaded data resides on the server as it provides a thin client web interface. In other words, no data is loaded on the client machine. Numbers are cached on the server but not text. The fact that CXAIR doesn't cache all the data means that available memory is not a restriction. Cache is used to improve the performance, but large internal memory is not a necessity but will help performance, particularly for ad-hoc calculations.
CXAIR is clearly a representative of a new generation of reporting/analysis tools that users can deploy to freely analyze data. It's a tool for self-service discovery and investigation of data. It's the tool that many data scientists have been waiting for and is worth checking out.
Posted December 10, 2013 6:52 AM
Permalink | 1 Comment |