Datascope Analytics is a data analytics and visualization agency in Chicago. Established in 2009 by Mike Stringer and Dean Malmgren—two PhD students in the lab of Luis Amaral, professor of chemical and biological engineering at Northwestern University. Mike and Dean were investigating large communication networks and scientific databases for information to support the lab’s research. Their realization that they actually enjoyed the data analysis, coupled with the growing demand for these skills, eventually led them to start Datascope Analytics.
I sat down with Dean Malmgren and discussed Datascope Analytics and the Chicago data visualization community in preparation for his presentation at the CDVG meeting on August 15. The notes from our conversation are posted here. He told me he is still a teacher at heart. This is evident by his passion for discussing data visualization and his work at Datascope Analytics. After reading this post, and hearing him speak at our meeting, I hope you are encouraged to reach out and speak with him. It may turn into an opportunity for you as Datascope Analytics is growing and has some exciting projects starting soon.
Your presentation for the August 15 meeting of the CDVG is titled “Data-driven: at the intersection of design and analytics”. Can you give me a little preview of what you will be speaking about and why you have chosen this topic?
I will walk through a project or two from start to finish to give a sense of how we approach problems and to emphasize the importance of designing compelling visuals to achieve our results.
Describe Datascope Analytics
We are a data-driven consulting and design firm. Instead of specializing solely in design, consulting or analytics, we operate in the space where these three functions intersect. With this broader perspective we are able to provide solutions customized for the data and challenges unique to our clients. We believe that this comprehensive approach has differentiated us in the analytics market.
What are some of your signature clients? Can you discuss the projects you did for them?
Proctor & Gamble contracted us because they needed their employees to adopt a new work process throughout a multinational and multi-functional organization. We conducted a social network analysis and created an influence network model. With this we identified the thought leaders and change agents who are simultaneously well-respected by their peers and optimally positioned in the influence network to foster a movement. The result was a team of ambassadors who will spearhead change in the organization.
We have developed a four stage process that has been successful for us. It has four major phases.
Clarifying our clients need. This is a collaborative exercise with the client to brainstorm ideas. We then select a few options that are the best fit for the problem and create prototypes. Our ability to create prototypes is one of our strengths.
Identifying the data that can be combined or created to provide insight for our clients. This may be data from within the client’s organization or from external sources. If the data is incomplete, we fill in the gaps with custom tools and surveys. All of this data is combined to provide a reusable asset for the client.
Designing the analysis. We know that our clients, and their data, are unique. Consequently, we don’t use the same analytical tools for every project. We are a custom shop because we believe we deliver greater insight in to a client’s data than we can with a vended software package.
Communicating the results with a LivingReport™. This is our unique solution that is more effective than just text or a table of numbers. It is a visual representation of the data that shows the patterns that can reveal valuable insight about the client’s business.
What technology does Datascope Analytics use to create their visualizations?
We use open source development tools to develop custom solutions for our clients instead of using vendor software. We feel the vendor packages have considerable functionality but are ultimately more limited that our custom solutions. For analysis, Datascope Analytics uses Python, R and Hadoop. Raphael.js and D3.js are used for visualization. We have also created our Lens™ library: a set of analysis tools that let clients see into their data with more clarity. It is built in Python and is the glue that sticks everything together coherently.
What is it like to run a data visualization startup?
My days are divided into three activities: white boarding solutions, coding, and interacting with clients. These aren’t eight hour days, however, so I get to spend a considerable amount of time on each of these. And that’s okay because I enjoy them all.
How would you describe the Chicago data visualization community?
I would like to see the Chicago data visualization community mirror the diversity of businesses that exist in our market. Unlike the financial focus of New York or the tech focus in San Francisco, Chicago has a very rich set of industries that could all benefit from data visualization excellence and cross-fertilization of ideas. This meetup one of several great ways to start the process of sharing ideas and bringing together the diverse community interested in data visualization.
What help in starting Datascope Analytics did you get from the Chicago community?
Northwestern University Farley Center was instrumental in getting us off the ground. They provided accounting services, space, and mentoring. We were also fortunate enough to receive deeply discounted legal assistance from the Loyola Law Clinic.
We have also benefitted from collaborations with other start-up companies in the Chicago area like Syndio Social.
Who are some of your favorite data visualization designers? What are some of your favorite data visualizations?
Moritz Stefaner and Stephanie Posavec are two of my favorites. I like how each of them thinks outside the box to come up with interesting ways of using different graphic elements to visualize data. Naming a favorite is difficult, but I particularly like Stephanie Posavec’s “11 x” series which, despite the simplicity behind the underlying visualization, is a fun way to explore the emergent patterns in the long multiplication.
What advice do you have for those interested in getting started in data visualization?
Get started playing with data any way you know how. Start with a pencil and paper, make a static image, and — if it is useful to do so — create something interactive. The only way to learn what works and what doesn’t is to try and iterate, not read and regurgitate