Tableau puts real-time data modeling in the dashboard
- By Brian Robinson
- Dec 01, 2014
Visualization has long been considered a prime tool for data analysis. But as data sets grow bigger and deeper and data management challenges mount, the need to make complex information accessible is paramount for government decision makers.
However, while current technologies such as Hadoop provide increasingly better ways to store and process data, methods for quickly turning that data into meaningful visual presentations have lagged.
Tableau Software’s interactive data visualization products offer an example of the next generation of tools that will allow just about anyone, across a broad range of skill levels, to produce those visualizations. Tableau takes that further, however, by also allowing near-real-time “what if” modeling and letting users share their analyses in a collaborative environment.
The City of Boston, for example, is using Tableau in a pilot program in the mayor’s office to set up dashboards for various departments in the city. If it works as envisioned, the mayor will be able to walk into a control center that contains screens showing the dashboards and not only get an idea of what’s happening in those departments at any given time, but also ask questions about particular indicators on the dashboards and get answers while standing in front of the screens.
The Department of Interior uses Tableau to improve its financial analysis, and specifically to find where it is spending its money and root out any discrepancies. The department also has a scorecard that includes “aging reports” that show transactions and payments over a period of time, according to Doug Glenn, deputy chief financial officer and director of the DOI’s Office of Financial Management, who added that Tableau helps to raise red flags about transactions that are getting old, “and can tell the who, when and why of those transactions.”
“For example, we had an accounts receivable problem where we saw the balance over a 180 day period was going up and we didn’t know why,” he said. “When we dived into it with Tableau we quickly saw the problem was with Fish and Wildlife Services, and that when we talked with them found it was due to Deepwater Horizon and British Petroleum.”
After the Deepwater Horizon drilling platform exploded and burned in the Gulf of Mexico in April 2010, eventually leaking nearly 5 million barrels of oil from a ruptured well head, owner British Petroleum agreed to pay billions of dollars to people who claimed they had been damaged by the spill. The Fish and Wildlife oversees that part of the program involved with natural resource damage assessments and reparations.
However, British Petroleum was holding up some payments out of concern that various agencies were putting questionable costs into reimbursement requests. Once the problem was highlighted with a Tableau analysis, BP got agreement for the requests to be validated before they paid.
Tableau Software, the company behind the application, includes four separate products that together to form a complete visualization platform. Tableau Desktop enables users to interact with Tableau to build their analyses, Tableau Server allows users to share and collaborate with the Tableau documents they’ve created, Tableau Online is the cloud hosted version and Tableau Public is a free product that lets anyone who has downloaded the desktop environment publish their data vizualizations to the web. The last feature is what Christine Carmichael, head of marketing, government and education for Tableau Software, tagged as “the YouTube of data.”
The products are based on two technologies developed by the company VizQL, which translates the drag and drop action users perform onscreen into database queries and then converts the response graphically; and a data engine that enables rapid, ad hoc analysis of large volumes data.
Tableau also has over 40 different connectors for most of the data repositories seen in the government and private sectors, and integrates with existing systems and security architectures.
While acknowledging that dashboards and visualization tools have been used in the government space for some years, Carmichael nonetheless claims Tableau goes further in terms of the intuitive interplay users have with the data.
“If you go back to the Boston example, think of the city having just experienced a nor’easter and now has 196 potholes that need fixing,” she said. “A year ago the mayor would have known he had 196 potholes, but then would have had to go to IT to get to the next level of where they were and how bad, and it could have taken days or weeks to get a response. With Tableau, the mayor can ask questions like that and get a response in real time.”
Teri Caswell, a senior associate with Hassett Willis & Company, works with government agencies to help them understand how operational, financial and performance data can influence decision-making. She’s currently working with first responder organizations such as the Federal Emergency Management Agency in a proof of concept program to see how rapid analysis of data can help them determine how things are progressing at disaster sites.
There are lots of tools that can do great visualizations and paint great pictures, she said, but there’s very few that can produce different visualizations with different data as quickly as Tableau can. She said it also allows her to work more closely even with clients who still prefer traditional methods of presenting data, such as spreadsheets and matrixes or text-based reports with visuals inserted in them.
With them, the “pretty Tableau pictures” are a bit of a hook that at least makes clients inclined to look at the visualizations, she admitted, but the real advantage is the ability it gives her to respond quickly and confidently to the iterative questions people ask her.
“Tableau means that, when I give a report and that next question is asked, instead of the usual ‘Let me get back to you with that’ I can instead immediately change a parameter or identify something that changes the visualization or numbers on the matrix,” she said, “At the least, I can tell them they don’t have the data to get the answer, and point them to what data they need.”
However, according to DOI’s Glenn, the days of traditional methods of analysis and reporting is coming to an end.
“The days of communicating with static spreadsheets or printed statements in reports is over,” he said. “The next phase is where we are dealing with large databases that might contain the answer, and the ability to drill down and sort, sift, analyze and manipulate data that exposes the answer no matter what the question is, is definitely the next step.”
Brian Robinson is a freelance technology writer for GCN.