overview

tags
UX Design, Data Visualization, Desktop, Web
time
Jan 2020 - Apr 2021 (Ongoing)
team
Office of the Executive VP for Research, Georgia Tech
role
The only designer
tools
Figma, Sketch, Excel, Power BI, Apple Pencil, Lucidchart, Miro

understanding

user needs
We interviewed one Associate VP and one Executive VP of Georgia Tech, and they wanted to make more informed decisions in managing research activities across Georgia Tech. I summarized specific questions driving the 2 major decisions.
In: How to increase research funds?
Regarding funders outside Georgia Tech:
1. How much, and how quick do funders offer? How has the situation changed over time?
2. Where are different funders’ specific interests, within Georgia Tech?
Out: How to spend research funds?
Regarding researchers within Georgia Tech:
1. How has research performance changed over time?
2. How are different colleges/schools/labs/individuals performing?
3. How are collaborations happen among colleges/schools/labs/individuals?

Other than answering questions in decision-making, users are also interested in exploring the data to raise better questions, as well as using the data as a way of communication to convince other decision makers.
know the data
We would visualize data of research projects initiated within Georgia Tech. The Enterprise Data Warehouse of Georgia Tech contains the funding amount, expected duration, and descriptive text of each individual project, as well as 3 other major categories of data.




1. Time and Status
During a certain period, a project may meet an action like: submitted, declined, awarded, or terminated.At a certain moment, a project may be in a status like: pending, unfunded, ongoing or expired. Further information can be explored regarding: How many days has a project remained pending? How many days has an ongoing project gone through?
By aggregating research projects of a certain action or status, the chronological data could reflect research performance.




2. Funders: Where does money come?
To help drive decisions in managing relationships with funders, I categorized and organized different layers of funding sources for each project.




3. Researcher: Where does money go?
To inform prioritization in research investment, I categorized and organized different layers of researchers for each project, as shown by this example of one individual under School of Interactive Computing, College of Computing.
comparative analysis
To gather inspirations for bridging such gaps between user needs and data, I studied some industrial solutions in visualizing research data for universities, as well as approaches of some other universities. For example:
Revenue Flow (University of New Mexico)
Research Awards$ (University of Virginia)

ideation

brainstorming
To solve users’ needs in analyzing, exploring, and communicating with the data, I exhausted major concepts of data visualization, to evaluate their pros and cons in presenting different facets of the data.
My conceptual graphs
My initial evaluations of all concepts
concept design
I selected some concepts based on initial evaluation, and then used Figma to mock up these concepts with data of our research projects.
user feedback
I collected feedback for all concepts by walking users through my Figma mock-ups.

However, there were so many alternatives in combining concepts with our data, and the initial feedback we got from users were not sufficient to inform my decision in prioritization among different concepts.

So we decided to have another feedback session. To get a better idea of what users expect from different visualization concepts, removed specific elements of our data and handed out 10 “raw concepts” with blanks for users to fill.
By revealing users’ expectation and preferences with these “raw concepts”, I started the design of actual dashboards by combining and prioritizing these visualization concepts.

design

Initial (low-fi)
I started with the simplest structure to meet user needs with 3 major parts.

“Activities” presented general data of research activities across Georgia Tech to inform decisions in prioritizing research funding. It showed how research performance changed over time, and how different units performed.

“Collaborations” helped decisions in how to spend research funds from the perspective of cross-unit collaborations.

“Funders” aimed to help manage funder relationships and increase funds, by presenting funding scale and interests of different funders.




Activities
As the landing page, “Overview” let users take a quick glance of the current status of research activities in the whole Georgia Tech, by showing key parameters and their YoY changes. Corresponding line graphs present MoM trends as well as YoY comparisons.

If users would like to see more details of certain information, “Units” allowed users to drill in and select units, projects and the variable of their interests. Users could drill further into certain units by clicking corresponding lines.




Collaborations
“All Units” visualized cross-unit collaborations by a network: Dots in a certain color represented schools/labs under a certain college, connections among circles represented research projects as collaborations among schools/labs. Users could click a dot to collapse it into multiple smaller dots representing individual researchers under the corresponding school/lab.

If users would like to see details of collaborations between selected units or individuals, “Projects” mapped collaborative projects as dots in a scatterplot matrix.




Funders
“Flows” presented sources and destinations of research funds. Users could click a funding destination to see how funds flow further towards units or individuals under it.

“Key Words” used word clouds to show funding interests of different funding sources. Users could click a box to dill into funders in a category of funding source.
mvp 1.0 (mid-fi)
Due to time constraints in development and database administration, I decided to focus on Activities and Funders-Flows, to make sure that the first Tableau prototype help with both aspects of users’ needs: researchers within Georgia Tech, and funders outside Georgia Tech.

I iterated the design to mid-fidelity, to help developers get started in preparing the back-end. Considering limitations in virtual communication caused by COVID-19, I added more details and annotations, to avoid ambiguity and misunderstanding.




Next steps:

- Collect user feedback regarding interactions(e.g. what to show on hovering?), to iterate the design to hi-fi.
- Adjust designs in case of problems occurring from development and databases.
please stay tuned…

this project will last for 1 more year. There is more to come!