Tableau Attracts Big Attention for Unique Approach to Data
November 6, 2012 Alex Woodie
One business intelligence software vendor that has been garnering big headlines lately is Tableau Software. The Seattle, Washington, company says it has come up with a better way for users to not only explore their large data sets, but to display the information in a visually compelling way. The software is being used with data from IBM i servers, but there’s more that could be done to streamline the connection, a Tableau manager says.
Tableau started as a research project of Chris Stolte, a graduate student at Stanford University who was researching and studying under professor Pat Hanrahan, who co-founded the animation company Pixar. The key technology underpinning the project and the company is VizQL (visual query language), which the company says enables users to query a variety of data sources (relational databases, Excel spreadsheets, or multi-dimensional cubes) in a visual way.
The big breakthrough in VizQL is this: Instead of viewing and selecting source data from rows and columns, as is traditionally done with SQL-based methods, VizQL presents the data source graphically, and allows the user to explore the data by dragging and dropping elements on the screen. This provides a much more intuitive user experience, and speeds the process of finding new insights and patterns in the data, and decreases training time and the need for expensive support from IT administrators, the company says.
“Essentially we make it extremely easy for any business user to connect to where their data is, and ask questions without needing to do any programming or scripting,” Tableau product manager Ted Wasserman tells IT Jungle. “It’s a very visual, rich user interface that enables you to ask questions of your data, while we take care of sending optimized queries or requests to wherever that data is.”
The Windows-based Tableau Desktop product lets users explore their data through dashboards that display a number of pre-defined graphs, ranging from simple bar graphs and maps to Gantt diagrams, histograms, and heat maps. Graphical depictions of data have always added some spice to otherwise boring reports filled with row upon row of data. But Tableau takes the reliance on graphics to a new level, and uses graphics to do new things.
“If you’re an average analyst looking at a billion-row spreadsheet, it’s kind of hard to detect on outliers and patterns,” Wasserman says. “So we take advantage of human perception to be able to display data visually. We use things like color, size, and special orientation to build and splice different dimensions [of data] together.”
Lately, Tableau has been feasting at the Big Data party. At the recent Strata Conference + Hadoop World 2012 event, the company announced new partnerships with Cirro, DataStax, Digital Reasoning, EMC Greenplum, Hadapt, Hortonworks, Karmasphere, and Simba. It already has partnerships with big data vendors Teradata Aster, Hewlett-Packard Vertica, IBM Netezza, and Actian Vectorwize.
Exploring big data sources–either in real time via direct connections, or via a data warehouse that is periodically refreshed with the latest data–is an area that Tableau sees a lot of promise. “In addition to just going back and viewing a dashboard over and over again, a lot of what people love about Tableau is the capability to do that ad hoc analysis,” Wasserman says. “Users can connect to a data set and look at it in different ways, trying to see what patterns there are–trying to understand the shape of data.”
Some IBM i shops also use Tableau as a front-end to query their DB2 for i (DB2/400) databases. One IBM i shop in particular, Ferrari and Maserati of North America, even participated in a Tableau case study, which is posted to Tableau’s website. In this case, Ferrari and Maserati of North America actually used HiT Software‘s DBMoto database connector to move the information between DB2/400 sources and Tableau Desktop running on a Windows PC.
Other IBM i shops have tried to connect Tableau with their databases, but with limited success, if the comments of IBM i professions on the forum section of the Tableau website are any indication. DB2/400 is not an officially supported data source for Tableau, with connectivity driven through a generic ODBC connector.
Wasserman admits that more could be done to help IBM i customers take advantage of Tableau, without relegating them to the vagaries of ODBC drivers. “It’s not something that’s currently on our roadmap [but] a lot of things affect the roadmap,” he says. If enough IBM i customers expressed an interest in using Tableau, “it certainly might make us go there in the future.”
The main challenge is utilizing the real-time connection between DB2/400 and Tableau. “If customers are using Tableau to query the data source live, that’s probably where they’re running into problems,” says Wasserman, who used to work in IBM’s DB2 group. “We do provide some capability to tweak Tableau’s behavior when using it with those data sources. It may be possible to figure out what tweaks to make and to make it behave better when connecting to iSeries.”
The IBM i market is peanuts for Tableau in the grand scheme of things, and the Tableau train will keep on rolling whether or not DB2/400 access is streamlined. Tableau as a company is growing at a rate greater than 100 percent, with revenues closing in on $100 million for 2012. Rave reviews from Gartner and its Magic Quadrant for business intelligence the last few years have helped drive customer adoption across all industries. Comparisons between Tableau and QlikTech haven’t hurt, and a Tableau IPO planned for 2013 will give QlikTech (which has a good number of IBM i and JD Edwards customers) some public competition.
As IBM i shops look for ways to make the most of their big data, Tableau Software is worth keeping an eye on.