Project Overview
DataKind is an organization that creates technological solutions for social issues. The HI Equity Tool by DataKind aims to reduce health disparities by identifying vulnerable communities and data gaps, providing social service organizations and researchers with insights to promote health equity.
I started this project during my master’s capstone at DataKind, which further evolved into a summer internship. As the sole product designer, I worked closely with the product owner and the development team to transform user needs into impactful design solutions, ensuring equitable access to critical services.
I started this project during my master’s capstone at DataKind, which further evolved into a summer internship. As the sole product designer, I worked closely with the product owner and the development team to transform user needs into impactful design solutions, ensuring equitable access to critical services.
my Role
Product Designer (sole)
Team
Product Owner
Product Manager
Product Designer
Software Developer
Product Manager
Product Designer
Software Developer
Duration
Jan 2024 -
Aug 2024
Aug 2024
Impact
- Adopted by 10+ partner organizations.
- Significantly simplified and quickened their workflow.
- Being continuously scaled and tested in multiple US states.
Problem
In the United States, millions of individuals and families face barriers to accessing essential social services, like food and healthcare, resulting in significant disparities in their wellbeing and quality of life.
Factors such as socioeconomic status, geographic location, and systemic inequalities contribute to these challenges, leaving many eligible individuals without the support they need.
There are many social service organizations, such as food banks and non-profits, that exist to help solve this issue. However, they find it challenging to focus their outreach efforts.
Talking to social organizations, I identified three main pain points they face
1
Scattered and Inaccessible Data
Data on community needs and social services exists but is fragmented across disparate systems.
2
Difficulty Identifying Areas and Populations in Need
Existing tools lack a comprehensive view of intersecting factors, making it hard to pinpoint areas needing targeted support
3
Time-Consuming Insights Generation
Creating actionable graphs to present to stakeholders for funding requests is slow and complex
1
Scattered and Inaccessible Data
Data on community needs and social services exists but is fragmented across disparate systems
2
Difficulty Identifying Areas and Populations in Need
Existing tools lack a comprehensive view of intersecting factors, making it hard to pinpoint areas needing targeted support
3
Time-Consuming Insights Generation
Creating actionable graphs to present to stakeholders for funding requests is slow and complex
And so the design question arises
How might we present actionable data about in-need areas in the US to help social service providers and social organizations make informed decisions?
Solution
Introducing the HI Equity Tool
All the relevant data in one place
- Consolidates diverse datasets into a unified platform.
- Allows users to overlay and analyze multiple datasets simultaneously.
Selecting datasets and topics the users want to work with
Adjusting indicators to get a focused view
Customize and focus on areas in need
- Maps data onto locations, uncovering patterns and correlations with a comprehensive view.
- Enables effortless customization of metrics and indicators.
- Allows users to focus on specific zones of interest.
Easily shareable insights and graphs
- Automatically generates graphs and charts for clear insights.
- Provides users with data-driven action recommendations.
- Enables users to quickly share insights and visualizations.
Insights and graphs generated for ease of understanding and sharing
So, how was the gap between understanding the problems and designing the solution bridged?
Research
Interviewing and observing a wide range of users
- To understand user workflows and challenges, I conducted in total 15 user interviews and contextual inquiry sessions with service providers, policymakers, and community advocates - the tool’s primary users.
- Observing users in their day-to-day tasks and talking to them revealed that users struggled with fragmented data sources and lengthy manual analysis processes, making it difficult to identify high-need areas and effectively prioritize outreach.
Understanding the current workflow and identifying the issues with it
Prioritizing functionalities that address the most major issues
- Building on the insights from user interviews and workflow analysis, I led focus groups with stakeholders, including service providers, developers, and data analysts, to collaboratively identify and prioritize features.
- Discussions focused on addressing their most critical pain point: time-consuming workflows.
- Using the MoSCoW framework, I prioritized features based on user needs and project feasibility, resulting in the following matrix.
For the first iteration of the tool, I decided to focus on the Must-haves and Should-haves. This prioritization not only addressed immediate user pain points but also provided valuable direction for future iterations of the tool.
ideation
The big challenge - consolidating all the functionality in a single tool
To ensure the tool is intuitive and user-friendly, I focused on three key principles
1
Mirroring familiar workflows
I designed the user flow to align with the existing workflow to minimize the learning curve and ensure seamless adoption.
2
Incorporating familiar layouts
I referenced layouts from tools users already interacted with, leveraging familiarity to reduce cognitive load.
3
Minimizing switching between multiple tools
The tool combines key functions - data selection, analysis, and visualization - into a single platform, eliminating the need to multiple tools.
Conducting a detailed evaluation of the various tools currently used by users, examining their design, and functionality.
Guided by these core principles, I designed the following streamlined user flow
The new workflow consolidates the functionality of at least 3 separate tools into a single, streamlined solution resulting in reduced complexity and time
design
Contemplating between different layouts
Initial Contenders
To refine the tool's layout, I explored multiple layout options through paper prototypes and quick usability feedback sessions. From this feedback, two distinct layout options emerged as contenders, each with its strengths and trade-offs.
Layout 1: Dashboard-Style Layout
+
Immediate access to all tools and options, flexibility for experienced users, efficient for multitasking.
-
Overwhelming for new users, visually cluttered, needs scroll bars for individual cards if the content is too long.
Layout 2: Step-by-Step Wizard
+
Simplifies onboarding, clear progression, minimizes errors.
-
Lacks big-picture context, time-consuming for simple tasks, need a lot of back and forth to refer to the map visualization and the insight charts.
Iterating on the initial layouts
After iterations and testing, a hybrid design was chosen, featuring a card-like selection panel for step-by-step guidance and a unified interface displaying all relevant information. This layout balances clear progression with a comprehensive overview.
testing and feedback
User testing sessions were conducted with the MVP of the tool, launched for use to the Second Harvest Food Bank in Florida. Participants included individuals from Second Harvest and other nonprofit organizations, such as United Way.
Some notes from the testing sessions
Key feedback and implemented changes
Making the search function more intuitive
Feedback: Separate search bars for datasets and location were confusing and redundant.
Change: I changed that to a unified search section, featuring an intuitive prompt like “Analyze [dataset] for [location],” simplifying the process.
Perfecting the slider design - leveraging proven design systems
Feedback: Sliders lacked precision and were confusing to use.
Before
After
Change: The slider design was improved by referencing established design systems, such as Shopify Polaris, since DataKind did not have its own design system. This approach ensured users could enter exact values directly for greater accuracy.
Making the homepage information accessible
Feedback: The homepage felt text-heavy, and lacked clear direction and CTAs.
Change: Text was minimized by using more graphic elements, and a demo section was added for easy access and understanding.
Before
After
Take another look at the final walkthrough
Outcomes & Learnings
A whole new domain: Jumping into the social service system with little prior knowledge was a huge challenge but also a great learning opportunity. Understanding community needs was initially overwhelming, but it helped me make more informed design decisions and showed me the importance of domain knowledge in creating effective solutions.
A new challenge: This project marked my first experience designing a large-scale 0-to-1 product. Starting from scratch meant building a vision and making sure every decision aligned with user needs and stakeholder goals. I also focused on detailed documentation to ensure the team could continue smoothly after my role ended.
A new challenge: This project marked my first experience designing a large-scale 0-to-1 product. Starting from scratch meant building a vision and making sure every decision aligned with user needs and stakeholder goals. I also focused on detailed documentation to ensure the team could continue smoothly after my role ended.
