2023-2024
When displaying the results of wind turbine inspections, it is necessary to combine high resolution photography with detailed data. As we began revising the way our damage and asset data was structured, it became clear that we needed to revisit the way we were displaying inspection results. Additionally, as we’d previously been handling photograph (both large format and thumbnails) in multiple ways throughout the application, I saw this as an opportunity to unify similar functionality in one, reusable package.
The first task was to identify all of the places where we were displaying inspection photography in the current application, along with all of the related workflows and use cases. I conducted multiple rounds of interviews with users to document their usage of the current system and engineers to understand what was going on under the hood. The intent here was two fold: one to establish a benchmark for minimum functionality, and two, to find current pain points and opportunities for product enhancement. After initial research was completed and processed, I had a list of top priorities:
- The full-size photography had to be given as much screen real estate as possible.
- The thumbnail photography had to have a robust set of features baked in.
- The interface had to be flexible enough to accommodate all the present data and functionality while giving us room to add more later.
- All the elements had to be modular enough to use and reuse throughout our product and portable enough to share with our sibling products.
After a few rounds of rough sketches and loose wireframes, I had a basic idea of what the final product was going to look like. For the thumbnail views, there would be a variety of features available that could be turned on and off based on the context they were used in. For the large image viewer, the image itself would take up the entire view, from edge to edge, top to bottom. All buttons and the collapsable data panels would float over top. I also identified places for future development, like enhanced navigation and location indication, that we would wait until to implement until we had more user data.
The team looked at the scope of the project and decided to focus on the Large Image Viewer first, as it was the piece we felt would require the most technical work and potentially the longest amount of time for user testing and iteration. I started working on a more complete set of lo-fi wireframes using a pseudo-Atomic Design Methodology approach, deliberately keeping the level of detail to a minimum. My goal was to figure out the required elements, how they related they related to each other, and roughly where they would go on screen. I have often found that adding more detail earlier in the process causes the designer and their collaborators to focus too much on aesthetics and less on functionality to the detriment of both.
With a set of initial pieces in my pocket, I began to play with the concept for the Large Image Viewer, adding higher fidelity as I worked, and gathering feedback from the team and our stakeholders. How do our existing workflows happen with this new set up? What about the new workflows we want to add later down the line? Where are those workflows clunky? Where are they confusing? These are the questions I asked myself and my teammates through five major iterations.
Those five major iterations were not the end of my work on the project. The more our Engineers worked on the solution and the more our Users test it, the more requests came in. Gathering continuous feedback, from even the earliest stages of development, and integrating it into our work made our product better and our solution stronger. When we were done, we had a feature that, among other things, could:
- Quickly load and display a succession of 61 mega pixel photographs
- Apply image filters to aid in analysis
- Mark and annotate blade damages
- Link damages across multiple photos and multiple inspections
- Quickly access information about the inspection and asset
- Record damage progression over time
Most importantly, the feature was well received by users and allowed us to unlock the potential of our new damage data structure to find insights in our customer’s data.
(additional process-related materials available upon request)
