%20Science%20of%20Designing%20Data%20Products.png)
Open your phone or your laptop, and you’ll realize that you interact with hundreds of data points daily. That data can include everything from your weather app displaying an hour-by-hour analysis of precipitation to the stock market updates in your brokerage account app. Data is everywhere, and our ability to trust and understand it allows us to make more informed decisions that influence our organizations and personal lives. Designing data-driven products and services is a team effort, and a key player on that team is the Data Product Designer.
In the context of civic technology, the users who interact with data products and services can be a staff member of an agency or the general public. Data users range from experts to beginners both in subject matter and data analysis. A Data Product Designer’s role is to understand their unique users and help them accomplish their goals with data as a tool. Data product designers don’t necessarily have backgrounds in statistics, but work with teams of experts including full-stack developers and data scientists.
Typically, the focus of communication for product design work is on the back and forth between designer and developers. With a data product, there’s another group to consider: data analysts including machine learning engineers and geospatial analysts who can work with tools from AWS, Tableau, Snowflake, Power BI, ArcGIS and D3. Designers connect users and data teams by translating complex, raw data into clear visuals and tools for better decision-making.
Designers work with analysts to communicate key data traits, such as accuracy and completeness. Case in point - election result visualizations on news sites will show the latest data available while using design elements to indicate if results are final or not yet confirmed. Showing users these data nuances can help them create contingency plans based on the potential impact of incomplete information. Misinterpreting data accuracy and completeness can lead to poor decision-making and negative consequences.
At this point you may be asking whether the product designer needs a background in statistics. It can be helpful to understand conversations, like whether to display the median or average of a dataset that doesn’t have a normal distribution. In reality, though, a statistics background isn’t a necessity, as long as the designer works closely with their stakeholders and teammates and asks critical questions along the way.
If your palms started sweating at the mention of “normal distribution”, the good news is your skills as a product designer still apply. At the end of the day, you’ll need to understand your user’s goals and work from there. Getting clarity on what your users need this data for and what decision it helps them make will guide you. What follows are some questions that will feel familiar and tools you’ll already have in your toolbox to help answer them.
Not all users looking at and working with data can be grouped neatly into a user type. In fact, many situations designing data products will involve end users who are not data experts at all but rather everyday people who need to interact with data to make decisions in their day-to-day work. The best thing you can do to design better data products is the same thing good designers do for any type of product—deeply understand one’s users.
Example questions: Are these data experts viewing your product or more typical workers? What sort of terminology might they use or understand? Do they work in this data every day or in an ad-hoc manner?
Design tool: UX research and personas
The data you’re designing around (hopefully) isn’t just there to look pretty or “interesting” but instead is used to achieve some outcome. Understanding the larger context of your tool will bring clarity to design decisions you’ll need to make and what features may be important beyond just data visualizations or tables.
Example questions: What standard is this data being held up to? Does it need to hold up in court? Do the numbers need to be exact to the nth digit or are the stakes lower and simplicity is key? How important is data provenance? Users say they need this data but does it actually tie back to user/business goals?
Design tools: Expert/SME interviews can help you unpack some of the more technical constraints and considerations. Stakeholder interviews with business leaders can help you craft objectives and key results that roll up into the larger mission.
Most likely, your data product isn’t intended to just sit on a screen, to be opened at the beginning of the day, stared at for eight hours, and closed at the end. Like other product design work, we have to concern ourselves and understand what our users are doing before and after they hit our product.
Example questions: Do users need to share the data? Do they need to share those findings with users outside their organization? What sort of permissions may various users need/have/not have? What other tools and programs do they use and what existing mental models might they have?
Design tools: Contextual interviews, user journey maps, service blueprints
Just like any well-functioning product team, the key to great design in this space comes down to clear and frequent communication - you’ve just added a new discipline into the mix.
The same good communication habits that enhance work with devs will go far when working with data analysts on your team. Involve them early and often and collaborate together on what’s possible. Working closely with data at the beginning of a project can be particularly helpful. Having a data analyst walk you through the dataset at the beginning of the project and share possible problems (for example, pointing out a field that isn’t reliably filled out) can save you headaches down the road. Involve data teammates with user research and usability testing too to ensure they see how users interact with your product.
To learn more about the design of data, books like Storytelling with Data by Cole Nussbaumer Knaflic, Show me the Numbers by Stephen Few and Envisioning Information by Edward Tufte can be excellent resources. If you have questions about designing data products, feel free to reach out to A1M Solutions—we’d love to share insights from our work in this space.
Ariel Yardeni is a Senior Policy Strategist with A1M Solutions. She is passionate about health equity and using research about the social determinants of health to influence policymaking.
We are dedicated to creating simple, sustainable solutions. Learn more about our Services.
