Industry Collaboration

I have a long history of industry consulting and collaboration with industry. A number of the visualization algorithms my team has published are featured in commercial software sold on the market today by AVL, Grid-Tools, VRVis, and We Predict Ltd. In addition, I have won a substantial amount of funding from our interdisciplinary work with companies including 3 PhD and 2 MRes studentships so far. My own PhD was funded jointly by two companies. The following are some of my industry projects.

Visual Analysis of Health Data with the NHS and Public Health England

The Full YouTube Playlist for this Collaboration

We have been developing visual analysis techniques for NHS and public health data since 2014. Our work studies complex, time-dependent, heterogeneous data sets including population health records including Clinical Commissioning Groups (CCGs) data, electronic health records (EHRs), hospital admissions data, and Alzheimer’s Disease Data from the The Alzheimer’s Disease Neuroimaging Initiative (ADNI). Some outcomes include:

• Using Clinical Commissioning Groups (CCGs) data to illustrate and compare the large prevalence of Hypertension and Diabetes throughout England,

• Showing the unique properties of the London region with better than average physical health but worse than average mental health.

• Illustrating the trends in hospital admissions over time with state-of-the-art visualization techniques for communication and storytelling with a wide audience,

• Visualizing the relative categories of both normal and emergency hospital admissions categories and those that incur the highest economic cost.

Visualizing the Customer Journey Through the Call Center Landscape with QPC Ltd.

The Full YouTube Playlist for this Collaboration

For the past ten years, call centers have experienced phenomenal growth in virtually every country around the world. The UK has approximately 1 million people employed in over 6,000 call centers. Some estimates say 3% of all UK workers were working in a call center (Holman, 2007). Estimates put the number call center agents at 2.9 million in the US and Canada with perhaps 17 million more scattered around the world.

Since 2014, my team has been developing novel, customized, state-of-the-art, interactive visualization techniques for call-center data together with QPC Ltd, a leader in call center software technology. We design visualizations that will empower the user by providing critical overviews of the call center data as starting points for exploration. We also provide zooming and filtering techniques that enable users to interactively select the sub-sets of the data they find most interesting. This visualization project also develops novel, state-of-the-art visualization techniques for portraying customer-centered details on demand. We draw on a wide knowledge-base of visualization expertise and current research literature to provide the most effective visualization solutions for the users including operations management, top level management, and call centre agents. Some outcomes of the consultation include:

• We showed QPC how and where they were losing customer calls in call centers.

• We derived a customer effort metric that models the customer journey through the call center landscape.

• We communicated and demonstrated our findings with novel visualization techniques.

Visualization of Engine Simulation Data with AVL GmbH

The Full YouTube Playlist for this Collaboration

I have a long history of scientific visualization as applied to CFD Simulation Data. AVL is a leader is motor engine design and development. The Department of Advanced Simulation Technologies (AST) relies heavily on engine simulation in order to reduce costs. The challenge of reducing time and costs for the product development cycle has created a growing demand to replace physical prototypes with virtual prototypes using simulation. Engineers face a great number of challenging development and simulation tasks which necessitate more than just good software. I developed a range of innovative, custom-made flow visualization solutions for AVL GmbH. The flow visualization solutions we developed for AVL-AST are incorporated directly into their commercial products which are sold on the market today. Our findings include:

• We identified and visualized critical areas of recirculating flow that prevented a cooling jacket from transferring heat away from the engine block.

• We tracked and illustrated deviant regions of fluid flow that prevent the ideal mix ratio of oxygen to fuel.

• We developed novel visualization algorithms for the visual analysis of unsteady flow and more.

Visual Assessment of Lending Risk with Oakbrook Ltd

Oakbrook is a digitally enabled consumer finance and technology business, removing the barriers between retail consumers that would like to borrow and investors that want to lend. Formed in 2011, Oakbrook brings together stakeholders with deep expertise in consumer lending, advanced analytics and forward-thinking technologists, with a shared belief that there is a better way to do consumer lending–by being more personalized and customer focused. Oakbrook develops software with a special formula for assessing lending risk for all loan applications. We collaborated with Oakbrook to develop visualization solutions for assessing lending risk. The software shows unusual gaps in financial transactions, outliers, in addition to typical behaviors. The visual solutions we developed resulted in a conference publication at the annual Eurovis Conference with the help of MSc students in computer science. Our findings include:

• We identified, using visualization, new areas of risk when analyzing customer data including travel, relocation, and multi-residency.

• We developed visual ways to depict customer spending and behavior including unusual behavior.

• We used state-of-the-art visual analysis techniques and applied them to the analysis of retail bank customer transactions.

Visualization of Automotive Warrantee Data with We Predict Ltd.

In order to successfully perform predictive analytics, assembling the right team of experts is critical. Our collaborative work joins the forces of practical, competent industry and curiosity-driven academia to form a synergetic symbioses of innovation. Formed in 2009, We Predict has grown rapidly with global customers supported through its worldwide network. We Predict exploits its global talent pool using modern, working practices. We Predict assists industry customers and partners in addressing critical business problems on a global scale. Predictive analytics can only be successfully delivered by people who understand mathematical and statistical principles that underlie modern data analysis techniques. My team introduced advanced, novel data visualization techniques into the process. Advanced data visualization methods are used to re-enforce, heighten, and extend the knowledge currently derived through the more traditional mathematics, statistics and computer science techniques.

We Predict have also hired three of the students that I have supervised in visualization. Outcomes include:

• We demonstrated how to visualize 3-way relationships and dependencies between automotive parts.

• We applied state-of-the-art and novel visualization techniques that are used to communicate our findings to a wider audience.