email us at Bioinformaticsolutions @ gmail . com

Saturday, June 23, 2012

Tableau - data analysis and visualisation made awesome


Another inspiring pair of presentations last week covering data analysis using Tableau, a new-ish (it's on version 7, so the analytics community have had their input into the development process) Business Intelligence product developed and commercialised by techies from MIT. First impressions? Tableau is awesome.

Having said that Google's BigQuery could save on the painful set-up and optimisation which facilitates data analysis of vast datasets using traditional database solutions - if you are prepared to or have already gone through the pain and are satisfied with the performance of your kit, this analytics tool will take care of the other pain of producing comprehensible and easily customisable analytics including any map-based representations you require as it has its own inbuilt WMS. The demos we saw using the Olympics travel disruption data and some healthcare financials were equally jaw-dropping - no coding required after you've pointed Tableau at your datasource (every conceivable variant of SQL database, or simple Excel and .csv files; online, offline, Tableau compresses extracts of big-ish datasets down to in-memory usable size - we wonder if you could, in fact, point it at BigQuery?), the whole demo was done using just a mouse, dragging, dropping and clicking - no function is meant to be more than six clicks away, most far lest than that - representing a massive time-saving.

And it's not just super-smart, it's pretty too - Stephen Few, the Visual Business Intelligence guru, helped to design the charting/representation options, and they really do create an impact.

You can find out more at both Concentra's blog where they demonstrate Tableau's web-plugin, showing off the mapping capabilities in regard of Clinical Commissioning Groups data and the Tableau home page - check out the free trial and also the various uses of Tableau server including Sharepoint integration.

No comments:

Post a Comment