Transforming how researchers and healthcare organizations bring data to life

Data Import

A medical imaging data management and analysis platform

Company

My Role

UX Designer

We face complex issues importing biomedical images and metadata into our platform quickly and efficiently. This is the first step to using our software, and it was a fragile, inconsistent, and confusing process for our users. Our research shows that 80% of a researcher/organization’s time is spent importing and curating data for downstream use. Our team was tasked to make this more accessible and less time-consuming.

The Challenge

Through data-driven research, we realized that we needed to double the volume of data our cloud customers manage in Flywheel, which includes making imports more accessible. Our goals are to:

  • Make the import user interface easy to use for any user

  • Update methods (4) for importing data to have parity

  • Make the workflow less fragile and less prone to erroring

  • Remove file size roadblocks during upload using auto-scaling

The Journey

  • High fidelity mockup of a dialog where a user can add data either from a storage provider or from their computer

    Add Data - Source Selection

  • High fidelity mockup where a user has uploaded files

    Add Data - Files

  • High fidelity mockup of a dialog where the user drag/drops or selects a folder

    Add Data - Folder

  • High fidelity mockup of a completed import dialog

    Add Data - Complete

As one of the first UX research efforts at the company, we used these methods to ensure we were working on solving the right problems related to importing data:

  • Conducted empathy mapping

  • Internal and external user interviews

  • Mapped the persona journey and workflows

  • Gathered qualitative data to drive interactions

  • Reviewing of previous requests

  • Usability Testing

Research Driven Process

The frustrations most mentioned in our research were the following:

  • Need for notifications

  • Waste of time

  • Fragility of our existing data upload methods

  • Inconsistent experiences depending on a user’s technical skill level

The Findings

What People Are Saying

“This 2.7 TB import went from taking 6 days, 4 hours in the old import process down to 1 day, 1 hour using the new import method.”

“… and without any handholding (without having to modify things for it to work). It’s really amazing that you can just transfer such an amount of data (and organize it according to Subjects and Sessions) using just the UI.”

— Pablo Velasco, Scientific Solutions Engineer

Impact

Results and Metric Gathering Ongoing

We are working on interviewing customers and watching metrics to gauge a level of success, including:

  • Ability to nearly double the volume of data our cloud customers could import (in bulk)

  • Total number of customers with >10+ imports

  • An increase in the average number of imports per customer

  • The success ratio for imports per customer trends positively

  • Increase in the total number of files successfully imported

  • Increase in the total number of bytes successfully imported