Neuro Omics Visual Analytics (NOVA)
Neuro Omics Visual Analytics (NOVA) is a data visualization tool developed to help neuroimaging and genetics researchers visualize and analyze their dataThis case study explores the design choices and user research behind the development of the tool.
UX Research
UI Design
Full-Stack
Researcher Pain Points and User Types
To understand the role that data visualization plays in neuroimaging and genetics research and the current pain points that researchers in this field face, I conducted interviews with current PhD students in my lab. I gathered several important insights:
- Data visualization is important because it helps reserachers identify trends and make conclusions based on experimental data.
- Many researchers who do not code very often struggle to generate graphs. The specific nature of neuroimaging and genetics research means that these graphs are often more complex or unique in format and therefore are more difficult to code.
Also based on interviews and scouring through the current literature, I discovered that there is a lack of comprehensive data in the field. There is great promise for discovering a more concrete relationship between our genes and brains, but the lack of data is limiting discoveries. Therefore, the pain points addressed are twofold: individual researcher struggles with generating graphs and lack of data in the field as a whole.
I sought to understand what types of researchers would potentially use a data visualization tool. There are several types of users I designed for, which have driven design choices and implementations:
- Non-technical users who are looking for an easy-to-use tool for visualizing data. This is the most important user I designed for based on insights from interviews.
- Discovery-driven users who are trying to discover new trends and exploring with a variety of inputs. These users require plenty of customization.
- Hypothesis-testing users who are testing specific hypotheses with specific inputs. These users also require plenty of customization and specificity in inputs.
- Gene-focused users who are looking for the impact of specific genes or traits.
- Multi-omics users who are seeking to integrate various omics data to understand complex biological interactions.
Researchers in genetics and neuroimaging tend to struggle with efficient and insightful data visualization because of their specialization in the field and lack of coding experience. There is also a lack of comprehensive data in the field. These factors are a problem because it creates bottlenecks in writing research papers and making meaningful discoveries.
Inspiration and Competitive Analysis
We decided that we wanted to develop an online tool that researchers could use to visualize and analyze their data. However, we were unclear about what types of graphs and visuals we should allow. I scanned through most available papers in the field with data visuals to find the most common visuals, which I found were manhattan plots and heatmaps.
I also looked into existing online tools that accomplished similar analyses. To see how useful they were in aiding researchers make discoveries, I developed a set of heuristics to determine the strengths and weaknesses of each tool. Most existing tools like TWAS hub and TWAS Atlas did not provide any strong or customizable data visualization features, which is what user interviews revealed was most important to researchers in terms of analyzing data. Only webTWAS provided a simple manhattan plot that lacked much interactivity.
Initial sketches of the website are shown below. I attempt to create a simple design that incorporates all the graphs and inputs that a researcher may need.
Prototyping and Iterating
The first goal was to build a manhattan plot. I decided to use Next.js as the web framework, and I used D3.js as the data visualization library. I chose D3.js because while it is one of the most customizable JavaScript libraries, which is necessary for more complex graphs like manhattan plots.
Below is the first iteration of the tool. There is a manhattan plot with customizable inputs and advanced filters that can be updated quickly. In addition, there is a table that contains the list of all genes visualized, which allows researchers who are interested in specific genes to learn more about them.
After, I implemented a second manhattan plot to accommodate a second dataset. This also allowed me to implement a heatmap for comparison across datasets for multi-omics users.
To accommodate researchers who are simply looking for simple visuals and comparisons, I also implemented a compact view that displays only the manhattan plots. This helps researchers make quick visual comparisons across two datasets.
TL;DR
Researchers in the genetics and neuroimaging field tend to struggle with generating meaningful visuals to help them make discoveries and identify trends. To address this issue and fill the gaps in the current literature, I developed a data visualization tool that allows researchers to visualize their data with common graphs and compare results across datasets. There are customizable inputs and filters to accommodate users interested in diving deep into their data, and the tool is designed for non-technical researchers to use. Below is the latest iteration.
