Actionable Data
Where Design and Data Science Converge
First, be warned. What I am about to share is vague and nebulous.
It is also extremely powerful.
Actionable data is tech-stack agnostic. It can be applied using any of multiple tool sets. It is a paradigm — which is to say, a way of thinking. It grows from the well-known discipline of design thinking, applying the first two steps: empathy and problem definition. From there it zeroes in on a particular solution set: designing a data narrative that applies data science principles in the context of dynamic, interactive user experience.
Most of us in 2020 are overwhelmed and inundated by information. We feel powerless and enfeebled by all the bad news. Actionable data emphasizes the power of informed collective action.
What actionable data does is take a measurable, quantifiable data set and connect it to specific calls to action. In essence, it presents you with new information and instead of leaving you to mull (or stew) it provides you with ways to respond. The goal is always to empower users.
In my online talk for Innovation Women, I discuss one case study in depth: The Giving Map, built as a COVID-19 relief tool.
Created in partnership with Charity Navigator, this interactive map is updated daily with the most recent data on COVID-19 infection rates across the United States. Users can click on counties and view case data, while making a donation to charity in any of ten giving categories. The tagline “Give in the areas hardest hit” led to a remarkable response. One out of 15 people who saw our posts on social media visited our site. This was with zero name recognition or major media coverage. If you know anything about Internet marketing, you know how unusual those numbers are.
User survey findings suggest that even when users chose not to give through the app, the availability of options contributed to a richer and more meaningful user experience.
This is just one example of actionable data as a specific application of design thinking: empathy with a user who feels disconnected and powerless, in response to the problem of massive social and economic disruption from COVID-19 in the U.S.A.
If you as a business, government, or nonprofit entity know of a problem that has quantifiable data associated with it (measuring either the problem or the response) then you have an excellent candidate to build a data narrative.
This does not have to be high tech. In my talk, I give the example of a restaurant manager who chooses to share online feedback with staff via text. Many applications of actionable data are possible that require no coding (or at most, a spreadsheet).
Actionable data is not:
- Cherry Picking — Picking one fact or statistic to make a rhetorical point and persuade users to take a given action.
- A Dashboard — Dashboards are great decision support tools, but they are typically developed to give a small group of users in a hierarchical organization multidimensional insights. By contrast, actionable data is designed to operate at scale. It is optimized for a many-user environment.
What is actionable data best for? Datasets that are:
- Dynamic — data is constantly refreshed.
- User generated — user input is captured by system variables to generate new data types and new levels of interaction.
- Too big to count — did you know most people cannot easily count more than five objects at one time?
Dynamic and user-generated content are wonderfully suited to actionable data — in particular, user-generated narratives are great for fundraising. But something as simple as a spreadsheet containing addresses of organizations in an informal cooperative network can, with a little massaging and some basic HTML, become actionable. Something as obvious as an interactive map with health advisories can and probably has saved lives over the past few days.
For Giving Map, we were able to build a custom API off the Johns Hopkins University Coronavirus datashare using Microsoft Azure, along with a brand new responsive web app, in just under two months. But if you don’t have the resources to spare for this type of rapid ramp-up, don’t despair.
Maybe you are a food pantry or a soup kitchen. If you have a way to share how many people were served with food (or how many were turned away) each day that data is actionable. You can use it as a tool to engage donors, volunteers, and other stakeholders.
Maybe your organization works with houseless and homeless populations in a large city. You may not have daily updates on your clients, but you can give visitors to your website a walk-through of different neighborhoods, sharing relevant facts and situational data, and giving visitors a way to connect at the neighborhood level.
This is more than a gimmick. There is a richness to data-driven UX and UI which compels further interaction and engagement. When data is abstract and featureless, it is easy to tune out. People engage at the level where data becomes specific and granular. The more you break down a given set of facts, the more you humanize it.
These are only a few hypothetical examples. We live our lives immersed in data. Thus, the data stories we can construct are nearly limitless.
To learn how to create your own map, visit our project repo on GitHub.
