Constraints
User stories were not finalised before design phase kickoff. Design team not involved during initial product discovery, and design research phase was not accounted for in the roadmap by the product team, leading to a challenging delivery schedule.
Early Wins
Competitive analysis on the other data analysis apps that the targets users were already familiar with helped drive fast decisions on the final product UI.
- Background
- The Challenge
- The Solution
- Impact
- Praise
[Client], the pharmaceutical wing of a Fortune 50 pharma company was looking to develop a Health Equity Software using open-source development framework, utilizing algorithms developed in-house by [Client].
The Health Equity Software is aimed to help healthcare providers and other users to perform guided analysis on their data, based on the integrated statistical algorithms and the Machine Learning (ML) models.
The client has developed ML/ statistical algorithms to support its vision of equitable health for all.
The proposed Health Equity Software will enable healthcare providers (hospitals, hospital groups, or clinics) or the scientific community to do guided analysis on their own data, utilizing statistical analysis and machine learning models developed by the client.
The Health Equity Software is intended to be deployed on the cloud and accessed primarily via a web browser.
Not only should the product help support researchers / clinicians make sense of data, but also made accessible to non-technical data analysts & students, both groups with lower exposure and experience working with healthcare data
Goals
1. User-Friendly Onboarding
Design an onboarding process that allows users, including those with non-technical backgrounds, to easily understand and navigate through the initial setup of the Health Equity Software.
2. Intuitive Navigation and User Guidance
Develop an intuitive user interface that facilitates easy navigation and provides contextual guidance to users throughout their interaction with the Health Equity Software.
3. Accessible Data Science Terminology
Enhance the accessibility of the Health Equity Software for non-technical or less-technical users by providing contextual explanations and definitions of data science terminologies
User Groups
- Admin
- Analytics User
For this Epic we were focusing on the Analytics user
Primary Analytics Users
- Clinical researchers / statisticians at hospitals
Secondary Analytics Users
- Data Scientists & Business users familiar with data
- College graduates (Bachelors) with STEM major
- People who want to work with either public or private healthcare data sets
What are the primary tasks and scenarios that the design should support?
Preliminary Analysis: Users have already finished building a data set and want to do some preliminary analysis just to see if there is some information contained in this data set, import this dataset into the system, have the software automatically run analysis & quickly visualize the results so they can better understand their dataset & their patients, and extract insights from it
Communication for Non-Technical Users: Design the system to help make it easier for technical users who want to communicate the plots & numbers with non-technical people
Healthcare Data Exploration: Add text or content to the results such as what the analysis indicates, what doesn’t it indicate, what are the conclusions & insights that can be extrapolated from this analysis. Make available Natural language description of these results
Design Principles
User-Centric Simplicity: Ensure that the user interface is intuitive, easy to navigate, and minimizes complexity in data analysis workflows. Simplify terminology, provide clear instructions, and offer tooltips to enhance the usability of the software for users with varying levels of technical expertise
Clarity and Interpretability of Visualisations: Design visualizations that are not only aesthetically pleasing but also convey information in a clear and understandable manner. Prioritize the use of intuitive charts and graphs, implement color coding judiciously, and provide annotations to explain complex data insights.
Guidance and Education: Integrate contextual explanations for data science terminologies, offer in-app tutorials, and provide proactive assistance during complex tasks. Foster a learning environment that empowers users to explore and extract meaningful insights from the Health Equity Software.
The entire project was divided into roughly three parts:
Module 0: Data Importing & Preprocessing
Module 1: SubGroup Analysis
Module 2: Model Unfairness Analysis & Mitigation
Task Flow Analysis
Click to View →
Key solution elements
- Wizard-style Onboarding
- Intuitive data manipulation
- Visual Data Representation
- Equation Builder
Design Exploration
After the task flow analysis, I focused on the following features:
- Enterprise Table design to make working with the medical data sets easy
- Data Visualisations to make the analysis results accessible & intuitive
Based on the insights gathered so far, we iterated a few initial wireframes for the homescreen.Once we had the go-ahead from the client, we crafted multiple UI designs for the homescreen, and presented them to the client.
Design Refinement
Our team crafted a few different approaches to the UI designs for the product based on the feedback we received & the subsequent discussions we had with the client's team. We tried to showcase different visual approaches to the user interface.
The client suggested to us that we don't spend too much me on the exact product illustrations / images yet as these are things that we can always change later once there are better suggestions.
Based on that feedback we decidedto put placeholder images where possible and focus on the usability, rather than exact visual assets.
As the UI team started crafting the design system and components for the homescreen, I crafted low-fidelity prototypes for the rest of the app. As we got the greenlight for each feature flow, it was passed on to the UI designers and presented to the client for review.
We updated the Module 2 'Health Disparity Analysis' wireframes & addressed the client comments from the workshops & the ones they left in Figma. We also revised & optimized the Equation Builder flow to make it user-friendly & also validated the same with our dev team.
"We really like the UI Designs & are looking forward to the MVP!"
- Client Team (Data Scientist)
"I just wanted to take the time to thank all of you for your efforts!
The demo looked really good – the tool is coming along really well - intuitive, well thought through and robust! I have high hopes of making this a useful tool to be used internally and perhaps externally. Thank you!"
- Client (Product Owner)
"Great to hear good feedback! Just had a look at the designs...looks refined & professional. Terrific work! Congratulations team!"
- Design Head
