Alex Miller saw the power of data visualization the morning after the 1993 Federal election when he looked at a map of the results. The president of ESRI Canada recalls, “I got a copy about 10 o’clock that morning and it was a jaw-dropper of a moment for me because you instantly saw that the ’93 election polarized the country. Quebec was purple. Ontario was red. The west was Reform green. In one instant you saw that this was a regional election.”
It took a remarkable effort for Elections Canada to produce that map 13 years ago, but today Miller only needs to open his laptop to show an animation of world population change. “It’s a shocking map. Look at Europe shrinking,” he said. “So which is going to be the superpower?” he added, as Europe turned a paler blue, and a band across the Indian sub-continent became a brighter thicker red.
Data visualization can produce dramatic insights from rows and columns and numbers. Today, leading organizations are making their Business Intelligence truly intelligent by displaying it in images that convey meaning – meaning that the business can use.
Sandy Struthers, CIO of Ontario’s Hydro One Networks, said, “It’s a phenomenal technology from the point of view that a picture is worth a thousand words. To be able to describe to somebody who knows little about the electricity system what it looks like and how complex it is and how we work is very valuable. It’s a phenomenal tool.”
With 1.3 million distribution customers spread across 640,000 square kilometres of service territory, Hydro One uses data visualization to pinpoint outages and quickly correct them. Today, the utility plots its people, equipment and assets visually; and in the future, the company will add graphic and video sources around outages as well.
“With the smart meters coming into use in the province, we will be able to identify when that outage occurs. We will be able to trace back against our system because we have a three-dimensional view of what our system looks like and we will be able to tell where the outages occurred,” Struthers said.
The output from data visualization software can be expressed in formats from colourful animations spanning decades of data to simple charts or executive ‘dashboards’, but it goes far beyond basic reporting. Even the simple displays are deeper, the data sources are more complex, and the interaction with users is richer and more interactive.
Christopher Ahlberg, CEO of Spotfire, Inc., a $50 million a year data visualization company, emphasized the dynamic difference. “What you want to do is bring live data and live analytics through visualization into the meeting. The whole event is transformed because now people from various domains can ask follow up questions in the meeting and bring all kinds of views to it,” he said. “A whole room full of follow-up questions – that’s where the real excitement starts.”
Who will step up?
Within an organization, there are probably as many paths to data visualization as there are potential users. Susan Terry, a Business Intelligence director with data warehouse firm Teradata, asked, “Who is going to yell within the organization that they need to test and try this first and who is going to step up to an ROI model that will let them do proof of value or proof of concept?”
Her colleague at the company, retail specialist Bill Franks, said the trail to value will be blazed by a champion. “It could be IT for networking, it could be a marketing person looking for customer analysis, it could be a merchandising person from a retail perspective. The key is, somebody is going to have to say, ‘I see value in this, I want to do it, let me take the lead in this pilot to demonstrate it.’ In the organizations we are talking to there is a huge range in who is stepping up to do that sponsorship.”
Business Intelligence firm Business Objects has an interest in creating more data visualization users, so this fall it released a freeware version of its Crystal Xcelsius visualization software. As Greg Wolfe, Business Vice President, explained, “For a long time, we gave away a free version of Crystal Reports and then we were able to upsell it for workgroup or enterprise use. We are trying that same formula with our Xcelsius offering.” With 3200 downloads in the first week, the ploy may work.
From Wolfe’s perspective, data visualization typically enters an organization through the business side rather than the IT side, “… except where you have quite a progressive and in-touch IT organization. When I think about the clients that are beginning to implement this, it has usually been a result of a line of business saying ‘we need a more visual approach to looking at some data in SAP’, for example.”
In some organizations, the demand for data visualization could start in the retail environment, where managers are flooded with point of sale data. Susan Terry said, “We are seeing things in terms of shelf space optimization, store space optimization, managing metrics and KPIs at a store operational level, down to managing different types of roles of employees on the floor with those metrics.”
Cost of implementation
Data visualization won’t come without a price. CIOs who implement solutions that incorporate large amounts of GIS data will face technical challenges of scale and performance.
As Sandy Struthers at Hydro One said, “From a data perspective we are putting large requirements on our data systems. We are storing large amounts of data on our system now, probably more so than we have ever done in prior periods. And we are also storing pictures, which takes up a lot of space.”
Computing power is getting less expensive and bandwidth is increasing, but simply scaling up resources may not be enough.
ESRI Canada’s Eric Miller said, “It helps us immensely that computing power is coming to us. But we can still bring the largest computers in the world to their knees, with no problem.”
The people challenge
For many organizations, Susan Terry believes, the challenge of Business Intelligence has been getting the analysis from the few to the many. “We still have the challenge of few power users – a little bit more in the mainstream and not enough in the casual space – actually banging away at the tools and bringing insights to their business,” she said. “Visualization is an opportunity… that will allow the bringing of that analysis from the few to the many.”
For the past ten years, Business Intelligence 1.0 has been about orchestrating data for one version of the truth, according to Christopher Ahlberg. Consistent, high-quality reports have been delivered to large audiences. People wanted organized and controlled data marts and warehouses, enabling predefined, frequent business questions and rapid responses.
“Now in 2.0, you don’t just have to enable one view of the truth, but many viewpoints that come together. And that is where it gets really exciting,” Ahlberg said.
The challenging part now for the CIO, he added, is figuring out how to take the next step and actually help the business compete. How can the CIO enable people to ask not just the predefined questions and not just generate nicely formatted reports once a week, but enable front-line people to ask any question? How do you get them working with more complex questions and answers?
To some extent, he added, people are meeting the need for visualization with a combination of available tools. “People do crazy things with Excel and Access at the front line. It happens nine times out of ten without IT’s involvement but how can you move enablement and ownership out to the front line where the real action is happening? That is where you really are creating differentiation. That is where we have the next ten years of challenges.”
Hydro One’s Sandy Struthers advised CIOs facing a demand for data visualization to understand how the line of business plans to use the applications and the tools. “They have to be along for the ride. It can’t be IT driving it,” he counselled. “You also have to have the right processes around data collection. The system is only as good as the data that exists in it.” To the extent that users don’t have good information, he said, the CIO will take the blame for it.
“The other thing we are very much pushing is ‘out of the box’ solutions rather than customized solutions,” Struthers said. “We encourage the business to align with the solution and then we push the ‘out of the box’ vendor in the direction we want them to go.”
Businesses that are information intensive are prime candidates for visualization. The Oil Patch as a good example. According to Spotfire’s Christopher Ahlberg, the company that wins in Alberta is the one that is smartest at taking the geophysics, the chemistry and the historical production data, and putting it together with financial data from the neighbouring field and some other information sources, then using all that information to guide the next $100 million drilling decision. That’s quite a picture to paint – and that’s what data visualization does best.
Richard Bray is a freelance writer based in Ottawa.