JR Reagan is the chief innovation officer for the federal practice at Deloitte Services. He teaches innovation and creativity at John Hopkins University and holds a masters in management information systems from Bowie State University and a BA in sociology from the State University of New York. In this interview he talks about Deloitte’s creative environment and how the company creates big data visualization. Furthermore, he talks about his vision of the future of the visualization space and open problems that need to be addressed.
Sramana Mitra: JR, let’s start by telling our readers about yourself and the program you run at Deloitte.
JR Reagan: I am a partner in our federal practice at Deloitte. I am our chief innovation officer for the federal practice, but I also run a unique center called HIVE (highly immersive visual environment). It is all about bringing the visual element to different problems – whether it is in the federal or the commercial world. My background is atypical and something one wouldn’t find in innovation or big data. I started in the military and did military intelligence for a while. Then I went to banking. I also did some startups. One of them went public back in the dot-com days. Then I went back into government consulting, having done consulting around the world to help governments solve lots of different problems, including cyber security or analytics.
SM: You said you didn’t follow a regular path. What do you mean by that?
JR: I think that a lot of folks grow up in a certain discipline, or they get into a career field and go through various corporate ladder steps to become successful. I had a different industry experience, and it wasn’t by design. I just found different opportunities in different places. I tell a lot of our young folks that I could not have gotten where I am today unless I had a bit of that zigzag path along the way. I am thinking about where I entered the innovation field. That was through a former boss who really took me under his wing and said, “From what I can see, what you are doing so far, JR, isn’t that innovative. Let me start putting you on a path where you learn what that is about.” Then I got the chance to study at Harvard and learn about disruptive innovation and similar things. I was very fortunate.
SM: What is happening in your program? What are you working on and what is exciting?
JR: For us it is a lot about visualization. I mentioned HIVE. Certainly the “V” in that really matters now. It is getting really tough – especially in the age of big data, when we are getting trillions of columns to make sense of – to see where the insights are. We are taking a different lens to that. We have applied design thinking into that approach to help our clients re-imagine problems and co-create solutions in a safe environment. For us, analytics and big data is certainly useful for algorithms and things that go on behind the scenes, but it is just as important to find unique ways to visualize problems and post them to the world. We spent 40 years digitizing data, and now we have to come to grips with how these problems collide – whether it is analytics, mobile, social, cloud, cyber, geo-spatial, etc. A problem isn’t a siloed analytics problem anymore. Our center helps people visualize and understand problems.
Sramana Mitra: Let’s take use cases of real customer problems and how you are using large-scale visualization to add value to those problems and generate solutions.
JR Reagan: One unique problem we thought was interesting comes from one of the large global retailers. Around the time of Hurricane Irene [in August 2011] they called us up on a Friday at noon and said, “What we need quickly is to have the ability to see what is happening from where our stores are located as well as where the hurricane is going to impact us in lots of different ways – where roads are being closed, where curfews are, where the supply chain is going to be impacted, etc.” We had to do this in a very short time. This was Friday noon. By Friday evening we had a prototype together. We talk about prototypes over presentation – having something quickly that we can put together and show to the client so he says, “That is exactly what I am talking about,” or, “No, let’s change that.” Then we were able to get a final version of that by 8 a.m. the next morning. It was an interesting problem, because it didn’t only involve the data they owned in their own four walls, but allowed data they didn’t own. We think that was a very interesting dynamic, because they have lots of tools and smart people, but they didn’t have the ability to visualize what they needed and do it quickly. That is one example of the kinds of problems we come up with.
SM: What are the data sources and what are the visualization tools and technologies you were using to get where you wanted to get to in this example?
JR: The data came in all different sizes and shapes. Those are near real-time data streams from social media, supply chain analytics, etc. Those really matter. If a store is impacted, there are lots of cascading effects to that. What do I need to bring there? Who does it impact? Where is it now? But they wanted to see it. They wanted to see a map – with layers. Depending on the question they wanted to answer, they could turn things on and off in order to see that. This case was a very geographic approach to layering on lots of different types of information. We use a lot of open source types of technologies. Traditional HTML5, JavaScript, etc. There are so many great widgets that have been created out there and we use those a lot. We used them, we layered them upon a map, we used a lot of open source geo-spatial APIs, and we created some of our own framework so we could bring together that layering aspect I mentioned before. On the back end, whether it is for this particular project or for other ones, we are plugging into things like Hadoop or Cassandra, as well as what we see from Academia. Things like D3 are being used also.
SM: What is the charter of your lab? Is solving government problems your main charter or, since you have innovation attached to the lab, does that also do a lot of internal innovation work?
JR: We started out primarily focused on the federal world. We were based around the government in Washington D.C. But it was interesting to observe the transformation of that. It wasn’t us evangelizing it. It was actually our clients talking to other clients. You have to go out there and experience this for yourself – there is this safe place to re-imagine your problems and figure things out. It turned out that we started getting just as many commercial clients – whether those were retailers, banks, or pharmaceutical companies.
Sramana Mitra: Your lab is now basically one that specializes in data visualization across different verticals.
JR Reagan: Yes. With innovation, we started getting called in to being more than just analytics. Now we are asked to come up with the next idea around the use of certain technologies in different ways – 3-D printing, for example. A lot of it revolves around analytics, but the innovation aspect comes in all different flavors, even in things like cyber security these days.
SM: Let’s do a few more use cases from other verticals.
JR: Another one I can think of is in the field of pharmaceuticals. A global pharmaceutical company came to us, and the problem they wanted to solve involved a combination of structured and unstructured data – a very big data format with lots of stuff. They would get complaints coming in to the call centers around the world, and these complaints would sit in a database – in a big text format and not organized in any way – and they didn’t have a way to look at them and ask, “Are there any trends in there? Is there any relevance we should care about other than somebody calling us and complaining?”
There is an example of where we used 3-Ds to visualize and mine the data. It was a very interesting visual that we presented so they could start dragging and dropping the hypothesis. “I wonder if I looked at this particular product, what are the complaints around that and who said it? Where was it?” What they found was that they didn’t know lots of the resulting insights. For example, for certain products, women in Asia would have a more adverse effect than men, or one particular manufacturing plant was having more defects than others. It was a very exploratory thing, where we would say that this data set and that visualization were giving answers to questions the client hadn’t asked yet. We were able to start drilling and mining through that. It had a big impact. Now they are planning to roll that out worldwide.
SM: Let’s do one more use case.
JR: Let’s talk about food and product safety. Using a combination of back-end text mining technologies, Hadoop and some visualization on the front end – 3-D and web technologies – we were able to put something together based on the hypothesis if there is a better way to look at food and product safety other than hearing about an event and then putting lots of people on a plane to find out what the cause of that was. It was very random initially, but it gets better as time goes on. We call it the Taco Demo, because it evolved around what would happen at a taco shack. What are the ingredients that go into a taco? Frankly, we had no idea there were so many. When you think about where the tomatoes come from, who distributed them, where they ended up, etc., and you know that taco shack number 4 had a food poisoning case, then they can start tracing that back.
It became a very relevant exercise for this particular government agency we are working with. It started to show that there is more ambient information out there that they could start looking at. People text all the time, “Hey, I am sick.” Hospitals are also collecting information. It also gave them the ability to start looking forward in terms of technologies. Products will come out and tell you the freshness of different products. But it also showed them that they needed to look at their processes differently, and that it may not be the most beneficial thing to immediately put people on planes. They needed to look at the problem first and then decide where to go. This is a great example of visualization and all the different things it can affect. Even though it might start out in the public sector world, it also has enormous effects to the private and commercial role, too. I love that story.
Sramana Mitra: What do you think is the state of the union vis-à-vis visualization technologies? I heard from several people that there are gaps in visualization technologies. You are saying you are primarily using open source technologies, and you have developed your own internal framework. Talk about what is driving those choices and what you are trying to accomplish.
JR Reagan: The good news is that we have seen a lot of tools, and the ability to use them getting democratized really fast. Tableau is a great example of one that people can learn pretty quickly and do amazing things with. That side of the house – the analysts, cheaper tools, the tools that can do things quicker – is progressing well.
Where we lack sophistication, because it is still hard and requires custom coding, is around the mash-up side of the house – when I want to create a geospatial map and layer on top some social media and some advanced analytics. That still takes a level of expertise. That is where we have to craft our own tools. But that is where the problems lie. We can start to see the light at the end of the tunnel, though, because we can see that everyone has a geospatial ability on their desktop now. That used to be the realm of experts. I think there are lots of different things appearing now wherein we will start to see a collision or a combination of those. Users, whether they are sophisticated or not, will be able to solve their own hypotheses of, “What if I layer these things on top, what does that produce?” even without a sophisticated knowledge of all the underlying data structures. That is where I think we are going.
SM: If you look at problems you would like to solve in the realm of visualization, where would you like to see entrepreneurs focus?
JR: One problem that is near and dear to my heart is cyber analytics and cyber security visualization. There are lots of different problems within that area. One is just scale that happens broadly across an organization. The ability to suck up all that information is an interesting thing. The second one is the speed at which cyber security stuff happens. It is really going to challenge big data to do the analytics on those edge points and quickly reveal those patterns and anomalies to act on them. It can’t remain how it is today, where we are just picking up the log files and shipping them over to some warehouse. By then it is hours, days, or weeks old, and it is not helpful.
Finally, there is the visualization aspect. What is the metaphor for cyber security going forward? It doesn’t always necessarily mean a map, because it is out there, it is a device traveling over a network. I heard somebody asking, “How do we visualize this when it used to be that the bad guys are out there and the good guys are in here, when all are on top of each other now?” The ability to do fast real-time at the edge analytics and the volume, velocity, and variety of that information is the next frontier.
Sramana Mitra: This is obviously a serious issue, and there are people working on it already. Who is working on it and who has good ideas?
JR Reagan: The good work is happening all over the place. You see the National Labs in the U.S. and other government organizations outside the U.S. that are doing some great work on the pattern types of stuff. There is serious analytics going on that is going to feed this once it is ready. Then there are the typical vendors out there that are starting to tune their platforms to accommodate big data in faster cycles. I would say a lot of great work around visualization and analytics is happening in places like Silicon Valley, academia, etc. But it is not necessarily geared toward cyber yet. That is what I think is going to be very interesting. Once people start to turn their heads to this approach or tool that I can use to solve cyber security, it will be an inflection point for us.
SM: How do you define cyber? The way the security world operates right now is that enterprises are protecting their own networks and infrastructure, governments are protecting their own networks and infrastructure, and individuals are supposed to have antiviruses and firewalls on their computers. How do you define cyber in this general area?
JR: It is going to be the collective risk posture of all of our networks and computers. It is a very broad statement on purpose. It used to be us all saying, “Wouldn’t it be great if we could visualize traffic jams and know where things are?” It used to be too hard. The technology was owned by somebody else, and they were sophisticated and expensive tools. Now you can go on your desktop and you can see it. We used to say that it was too hard to figure out where a disease [was most active]. We can do that today. I think it is about asking a different question. People are solving it within their four walls, but when we have this hyper-connected risk, wouldn’t it be great if we could see where bad things are happening on networks across the U.S.?
SM: You are talking about a public good that would be monitored by private data providers that are able to provide information about exceptions, intrusions, or malware to various bodies that are in a position to act upon it.
JR: Absolutely. You could even envision a situation like this one: “My connection in the house is slow. Check on that. Maybe it is a denial of service attack.” You have to have an overreaching hypothesis and vision to really stretch where this will go, not just solving one incident after another. I think that is where we are at right now.
SM: Tell us a bit more about your lab. How is it structured and how does it operate?
JR: We have a couple of parts to it. We found that the experience in that lab is almost the most important thing. It is about prototypes, not presentations. We train folks to be client experience specialists. It is like in a big theme park. They know how to really engage people and get them to have great conversations about what we do. So we have those people, who are experts who can dive into various problems, and we also have what we call Visualization Studio. We don’t typically find that in IT organizations. You normally find those on Madison Avenue [in advertising].
We found that using design-led solutions allows us to get over the hump of understanding if it is the right problem to solve or not. We have designers, engineers, data specialists – we cohort them on each problem. Those are the two worlds that make up our lab. Last, we have something you call a GovLab, which researches big problems. We do that through digital native eyes, so they can look at problems way differently from people like me, who have been around the block and seem to know it all.
Sramana Mitra: What is the source of the problems? Is it the customers who bring the problems to you, or do you bring the problems to the customers?
JR Reagan: We normally have “Art of the Possible” sessions – they are about two hours long. Sometimes [people] are very eager and want to launch off into something, and sometimes they don’t really know. They have questions like, “I wonder how I would optimize my supply chain?” Those are very general things, so we start walking them through a series of prototypes and visualizations. Normally what we see is that at a 15-minute mark, they start either getting up and touching the screen themselves or almost grabbing at the air and describing the solution right in front of you. It is phenomenal to watch. That is where they start bringing the problem to us. We had some clients to leave several pages of ideas they wanted to tackle. But we also have scenarios we call BYOD (bring your own device) versus bring your own data.
SM: So, they give you data sets and see what kind of insights you can gather out of them.
JR: Exactly. I was always wondering why they would do that. It turns out they don’t have designers, usually. The creative side is missing in a lot of organizations.
SM: When you have identified a problem with a customer, you have an understanding of how you want to solve that problem, what kinds of data sets they are bringing in and what of visualizations you are going to do, do you go to other possible takers for that particular solution?
JR: Yes. I can’t tell you how many prototypes we make, show them to somebody and they would say, “Wow. We could use that.” And they would tell me how they could use that in some other industry I had never thought of. They just find themselves in these visualizations. I shouldn’t be amazed anymore, but I am on how that works and how quickly people just need to see something and react to it, because they can’t come up with it on their own.
SM: Do you ever experience people in your organizations wanting to go off and create their own companies? When you see problems up close, sometimes you see the potential of taking that problem and building a business out of it. Do you see people wanting to do that?
JR: Yes. I mentioned the emergency crisis scenario on hurricane Irene. That spawned into one where we had many organizations say,“We would love for you to offer that to me as a service.” Those kinds of things started leading us to think whether or not we should start building more types of businesses around that. Those are the kinds of things we are exploring.
SM: That is very interesting. Where is your lab?
JR: The lab is in Rosslyn, Virginia.
SM: And how many people do you have?
JR: We have 20 to 30 people. It grows and slows on any given day.
SM: And it is all in one place?
JR: We have other labs around the globe in our member firms, and we just finished adding a group of them from all over the world and we train them. This will become much more of a global scenario for us.
SM: Thank you for sharing your story with us.
JR: I enjoyed it.