Roambi Blink: In-memory analytics and data discovery on iOS devices
Underlying the design (in both graphical and engineering senses) of Roambi Analytics is what Alsbury calls "in-the-moment usage" -- that is, short task-oriented bursts of data interaction by people working in sales, factories, retail outlets or field service, for example. The result? "Views and visualisations that are designed to fluidly guide the user through a dataset...all built around native applications where all of the visualisations and rendering live inside the app, as well as all of the data, which is brought down and stored locally on the device, so there's no latency when the user is interacting with it and they can use it when they're offline."
Roambi's customers are predominantly large enterprises who push data out to "massive numbers of mobile users, in pretty much every industry you can imagine," says Alsbury. Roambi Analytics originally used an on-premise server to connect to customers' databases and business intelligence (BI) systems, but a hosted cloud version was added around 18 months ago in order to attract smaller companies lacking the IT infrastructure and skills to deploy the on-premise server.
Blink
This move to broaden the potential customer base underlies Blink, a free add-on for Roambi Analytics 8.0, which is unveiled today. Currently, unless you sign up for one of the subscription packages, you won't be able to do more with the Roambi Analytics app than explore a few sample views. The idea with Blink, says Alsbury, is "essentially to bring a Dropbox-like experience to analytics, to every mobile user, for free, on their device."
Blink includes a new visualisation that enables what Roambi calls 'freestyle analytics', which is the ability to take any data set and, on your iOS device (there's no Android or Windows Phone support yet), instantly explore and analyse it, with no configuration required. Underpinning that visualisation is an in-house-developed, patent-pending, in-memory analytics engine designed to run on a mobile device, with no connectivity to a server required. Blink will be compatible with iOS devices from the iPad 2 and iPhone 4s onwards.
With Blink, you can open CSV files on your iPhone or iPad from email, or cloud storage services like Box, Dropbox, Google Drive or iCloud Drive, and instantly begin filtering, pivoting, exploring and sharing data visualizations with other Roambi Analytics users. In his demo with ZDNet, Alsbury first used a basketball-related dataset containing 90,000 rows within which Blink automatically discovered the various metrics and a time dimension to produce an initial chart, which he then filtered and pivoted on an iPad to extract different insights, noting the "mind-boggling" absence of latency as he did so. The automatic data mapping can be edited, and new customised metrics added as required. Individual snippets of data can be highlighted and sent via email as screenshots, or the entire dataset can be shared with other Roambi users via an email link.
"We've been working on it for two years," says Alsbury of the in-memory analytics engine that powers Blink. His second demo was perhaps even more impressive, manipulating a 30MB dataset containing the box-office performance of every movie released between 1942 and 2015, by month and year (time dimensions that Blink automatically detected and presented as visualisation controls) on an iPhone -- again with no discernible lag.
The value of Blink, says Alsbury, is the way the new analytics engine -- which comprises a proprietary columnar database for fast in-memory analytics, a parsing and heuristics component for no-configuration data view mapping, and the data model for pivoting and viewing the data from multiple angles -- enables what Roambi calls 'distributed analytics': "all the server is doing is distributing the initial data set to the end user, and then the device does all the compute -- it's an extremely efficient way to leverage the compute power that's sitting out there, latent in all of these mobile devices."
A new free Roambi account will allow anyone to download the Roambi Analytics iOS app, access CSV files from anywhere (other file types will be supported in due course, says Alsbury), and explore, visualise and share the data. There's also a new entry-level paid-for subscription, Starter, which supports up to 10 users for $10/user/month to complement the existing Business and Enterprise plans.