CellProfiler Analyst Web
UX DESIGN 2021
Web based cell classification software I helped redesign for the Broad Institute of MIT and Harvard, with the effort to make the tool more accessible and easier to use.
My Role:
User research, prototyping, UI redesign, and academic writing
Team:
Yahiya Hussain(machine learning), Emma Kelminson(Developer),
Thouis Ray Jones(Client),
Loriane Franke(Academic advisor),
& Daniel Haehn (Academic advisor
Challenge
CellProfiler Analyst (CPA) desktop software has enabled the scientific research community to explore image-based data and classify complex biological phenotypes through an interactive user interface since its release in 2008. However desktop versions of CellProfiler Analyst have been prone to difficult and time-consuming installation and version management issues. The goal of redesigned web version of the software, CellProfiler Analyst Web (CPAW) was to allow for greater accessibility, quicker setup, and facilitates a simple workflow for users.
Requirements for redesign established in Pre-design user interview:
-
Client-side only
-
Similar functionality to existing software
-
Easy to use
-
Must utilize CellProfiler data as input
-
Must output reproducible training data for further use
-
Interactive and reactive to user input regardless of background computations
-
As preferment in machine learning pipeline as original software
CellProfiler Analyst Desktop UI
Redesigned CellProfiler Analyst Web UI
React.js and Material.ui provide a clean and modern designed user interface look in the browser. CellProfiler Analyst Web is accessible without the need of any installation, while CellProfiler Analyst 3.0 requires installation which can be difficult for some users.
-
Four domain experts from The Broad Institute of MIT and Harvard using CellProfiler Analyst for their research
- 2 male and 2 female
-
30 minute semi-structured interviews via Zoom meeting
-
Experts asked to complete tasks that demonstrated the core classification functionality of CellProfiler Analyst.
- Half used CPA and half used CPAW to complete the tasks
Quantitive results:
-
Expert’s were able to complete all essential classification tasks using both tools.
-
Users completed their assigned tasks using CPAW 20.86% faster than with CPA 3.0
- 5.67 minutes vs. 7.17 minutes
Usability Study
Qualitative feedback:
Overall, the expert users were enthusiastic about the web version of the software, especially when considering new users and researchers using CPA for short-term projects in the future. One of the experts stated, “the biggest appeal is having it easier for new users to use”, and that “installing CPA was sometimes not easy.”. Another expert expressed that “it’s definitely convenient to use if you don’t want to download and install CPA itself”.