Prison Data Visualizations — Advocating for NY
OVERVIEW
Timeline
Fall 2020, 6 weeks with Cornell Hack4Impact
Team
2 Designers, 2 Project Managers, 5 Engineers
Objective
Create a Data Visualization Platform for the Correctional Association of New York to track updated prison data.
CONTEXT
Problem Statement
Currently, there is a lack of transparency in New York about the conditions inside prisons. The Correctional Association of New York (CANY) is the only independent organization in New York with the authority to monitor prisons and report its findings. To increase transparency of this data, Hack4Impact is working with CANY to create a data visualization dashboard that can be used to understand and keep track of current prison data.
My Role
I joined this project after user research had been conducted and a tentative information architecture had been determined. As a team, I worked with Connie Liu, our other designer to work on interaction design, visual design, and building the final prototypes.
PROCESS
Information Architecture
Through talking with our clients, we learned more about how CANY uses data in their reports and what kind of tools they needed within the data visualization platform to manipulate the data.
Based on this, we created an information architecture model where we divided the types of tools needed into three groups. The first group edited the time frame and type of graph being displayed, the second set changed the data set inputted into the system, and the third set filtered out specific data to look at within a data set.
Low-Fidelity Sketches
After determining the information architecture, we worked on creating low-fidelity sketches to try out different layouts and features for the data visualization dashboard.
The low fidelity sketches revealed several product specifications to sort our with the client such as whether there will be a default graph already set up when users visit the dashboard and if users need a way of saving and accessing previously created visualizations.
Mid-Fidelity & High-Fidelity Sketches
After clarifications with the client, we turned these sketches into mid-fidelity models.
As we were working with the mid-fidelity sketches, I wanted to improve the overall layout for the dashboard. I felt that from the user perspective some aspects such as the summary of the current graph settings bottom left were confusing and that it might be difficult for users to figure out which tab to click on to filter the data set without having used the dashboard.
Based on this, my partner and I decided to work on creating two high-fidelity sketches where she continued to build on the current layout and I explored a second layout.
In Layout 2, I made a few different changes. Instead of having a bar at the top with tabs that change the side-bar settings, I created a drop-down menu to make it easier to quickly change between settings within different tabs (the user can open all of the drop-down menus). I also eliminated the summary box at the bottom left-hand corner which showed a summary of all of the settings that user had selected to create the graph, instead choosing to place the summary under each drop-down option. This was done to help a new user navigate to the correct tab when trying to change the setting.
When presenting both designs to the team, we consistently received the feedback that Layout 1 looked more professional and less cluttered than Layout 2. Based on this, I revised the design for Layout 2 by changing the typography and button layout to create a second version for user testing.
User Testing
After completing the high-fidelity sketches, we tested the designs on both CANY interns (who will be the primary users of the dashboard) and others. Across all of our interviews, users found the drop-down style of Layout 2 easier to use, but the minimal interface of Layout 1 to be visually more appealing.
Based on this feedback, we decided to proceed with Layout 2 for our final designs, but worked on improving the visual design by creating a clear sense of hierarchy.
For example, we changed the summary text under each drop-down tab to a gray and added a vertical bar to the side of the text to make it more clear that the text is a subcategory within each tab that provides further explanation.
Revising the Information Architecture
Before finalizing the designs, we met with our client again and realized that our Information Architecture did not fit with the expectations for how the dashboard would be used.
While simpler to use, the original information architecture model assumed that the user would not want to place multiple filters at one time. For example, if looking at a data set that showed number of complaints from prisoners by county, the user could not select and look at multiple counties at the same time. They would either have to look at the data for one county or all of the counties. This was quite limiting.
In order to increase the flexibility of our platform, we created a second version of the information architecture which follows the same structure as before, but allows for multiple selections in the second two groups:
In order to incorporate this new structure to the dashboard, we included an additional pull out tab that allows the user to submit multiple filters for the data set.
FINAL DESIGNS
Currently, we’re in the process of building the finalized Data Visualization Dashboard. In the meantime, here are some of the prototypes we created:
Over the next few weeks, our team of engineers will finish creating the data visualizations and building the platform for CANY.
