THE PROBLEM:

ACME is a global decision analytics company that has worked with some of the world’s largest credit card companies to screen transactions and highlight suspicious activity using its proprietary EMCA software.* The current tools sometimes provide inaccurate forecasting models, and it can be difficult for analysts to identify the “big picture” or receive timely feedback on solutions. The focus of this design challenge was to improve the user experience for Jane, a risk analyst who is monitoring the performance of a current fraud strategy and taking actions to improve that strategy if needed.

This project was coursework for SJSU’s ISE220: Interaction Design II, Fall 2021.

*To protect the identity of the client, “ACME” and “EMCA” are codenames for the project during the visioning phase and will not be used in-market. The designs, content, and experience portrayed are purely conceptual and not guaranteed to be built in their full actuality.

MY ROLE:

Designer

I worked independently to design a high-fidelity interactive prototype over six weeks.

Strategy, Scope & Structure:

This design project started with a comprehensive review of the tool and data requirements outlined in the brief. Risk analysis and fraud detection were new topics for me, so I also did some research online to better understand the terminology, and the user’s goals and needs. I was then able to build out the site grammar and a priority matrix for Jane’s persona.

Jane needs a dashboard with high-level performance information, and ways to explore the two most important objects: rules and hotlists. These three elements formed the basic structure of the site.

Data visualization:

An important part of the conceptual model was outlining the most important measures, and their respective dimensions and timeframes. Jane needs to quickly and correctly identify fraudulent transactions to save the company money without annoying customers, so her key performance indicators were revenue, rule accuracy (correct identification) and precision (correctly identified fraud), and false-positive rate (percentage of legitimate transactions incorrectly flagged as fraud). There were also important measures related to the rules and rulesets, as well as hotlists (groups of items associated with suspect activity).

After defining these measures, I constructed an information architecture to define the word design (e.g., tone, voice), and search/browse designs (filter, sort, zones, etc.) that would help Jane find information quickly.

The final design integrates the four critical measures with data visualizations that either compare, categorize or normalize the information in ways that help Jane see the ‘big picture’ quickly on a single screen. Throughout the platform, I used bar charts to depict differences (e.g., between rules and rulesets), line charts to show performance trends over time (e.g., transaction volume), and combination sparkline + bullet graphs for space-saving views of trends and comparisons (e.g., individual rules).

interactive prototype:

My EMCA Fraud Analytics interactive prototype was built in Axure. The site leverages microinteractions that let Jane search, filter, sort, and re-visualize data, highlighting and brushing to keep multiple data visualizations in sync, mouse hover details-on-demand, and the ability to change the timeframe and reference for comparison. A screen capture of the dashboard is shown in Figure 1.

wrap up:

What I learned from this project:

• The goal of data visualization is to shift the power balance from cognition to perception - How can we tell the story at a glance?

• Most of the time, data visualizations can be solved using big numbers with trend arrows, and a clear visual hierarchy.

• To show differences, use a bar chart; to show trends, use a line chart.

• Humans are bad at visual estimation so avoid using pie charts (angles!), stacked bar/area charts (math!), or gauges (unless they’re clearly marked and show real-time fluctuation).