Last updated: March 3, 2019
If you are interested in doing research under my supervision, this page lets you familiarize yourself with my work, expectations, and general lab practices. While it’s primarily geared for Ph.D. students, there’s also information for undergrads and masters students.
What types of research can do you?
The first thing to know is that my lab researches information visualization, visual analytics, and human-computer interaction (HCI). Data Vis and HCI are some of the most dynamic and exciting research areas within Computer Science. The mind map below shows some of the concepts and techniques that are frequently synthesized in our research—it’s a broad field that exposes you to many domains. (Also see my Research page for a more in-depth description of Vis/HCI research.)
If you want some other perspectives, these posts give an excellent overview of what it actually means to research visualization:
The conferences that we publish at include IEEE Vis, EuroVis, PacificVis, CHI, and CSCW. Looking at recent work from those conferences will give you a good overview of the current trends in the field.
What makes for a successful Ph.D.?
There are many terms you commonly hear that describe successful Ph.D. students: hard-working, self-motivated, team-player, passionate about research, etc. These are all true, especially if you want to be a member in my lab! But I also want to mention one important factor in successful Ph.D. students that I often think gets overlooked: “the grind.”
Finishing a Ph.D. while keeping your sanity is all about the grind. Instead of relying on code sprints, last minute all-nighters, or intermittently starting-and-stopping your work, you should consistently apply pressure towards a goal (attaining a Ph.D.) over a long period of time (4-6 years) by maintaining a disciplined approach to completing research projects. More than anything, that is what makes for successful Ph.D. candidates and prevents issues like burnout and exhaustion. If that doesn’t sound enticing to you, you’re probably not a good fit in our lab.
As an example of what the grind might look like on a Data Vis or HCI research project, you might do the following tasks: plan out and design a project, code a prototype system, conduct a user study to evaluate it, statistically analyze the study results, and write the paper describing the process. Then the fun part: you attend a conference and present your work in front of some of the top scientists in the world. Especially early on, my mentoring approach focuses on making you a successful grinder. By the end of your Ph.D., not only are you a top expert on your dissertation topic, you are a well-rounded scientist capable of overseeing the full research project lifecycle.
For some other links on what it’s like to be a Ph.D. student, I recommend these:
- The Ph.D. grind and Advice for early-stage Ph.D. students, by Philip Guo
- Modest advice for graduate students, by Stephen C. Stearn
- How to survive a Ph.D.: 22 tips, by James Arvanitakis
- Choose your own indenture: You are a prospective grad student, by Adam Ruben (funny)
Joining the Lab
- Ph.D. students
- If you would like me to consider you, please note so on your graduate application to ASU. Feel free to reach out via email after you have applied. In general, having an online presence gives you a competitive advantage: particularly if you have (1) a LinkedIn profile and (2) a personal website that contains important academic information like published papers (with links to PDFs), a CV, contact information, a research statement, and brief descriptions of prior research/software/projects. It’s perfectly fine to link these in your application’s statement of purpose or CV.
- In your first semester, you are expected to take the Data Visualization gradaute course (CSE 578) and do well in it. The course is designed to be research-focused, so it’s a good primer for future research and allows us to get to know each other better. If you do poorly in CSE 578 or we realize there is a bad student-advisor fit, it’s better for both of us if you switch to another lab. If you’re research includes topics like VR or ML, take the corresponding classes ASAP.
- MS Students
- If you are interested in doing an MS thesis, you must first take CSE 578. During the semester, approach me and we’ll have an initial conversation. We’ll review your work at the end of the class and make a decision then.
- Undergraduate Students
- If you want to do visualization research for your undergraduate thesis, email me with a brief description of what you’d like to work on and we’ll set up a meeting. You should come prepared to present a high-level outline of what type of work you’d like to do, what you think the inherent challenges will be, and a rough timeline describing how your goal can be accomplished.
- If you are just wanting to get research experience as an undergraduate, I am also happy to let you join a graduate student’s project if I have an opening. I usually cannot provide funding for undergraduates, though I am open to letting you conduct research for course credit. In addition, the FURI program at ASU (which you should certainly apply for!) provides a nice stipend for doing research.
Mentoring: I believe that, especially early on, clear communication is extremely important from the advisor to the Ph.D. student. New students in a research field almost always need guidance, especially about what constitutes good publishable research and how to plan and successfully execute research projects. A good analogy is bicycle training wheels; they prevent you from falling over while you’re learning to ride on your own. If you keep falling over, you’ll quickly get discouraged and not want to continue (and your productivity will plummet). When starting out, you’ll be closely mentored either by myself or a senior student. As you gain experience, becoming proficient at conducting projects and publishing papers, you’ll begin to take a leadership role in the lab, mentoring and helping younger students. Depending on your post-graduation goals (get an industry job, become a professor, work in a research lab, etc.), I will work to put you in a position to succeed, such as coordinating internships and networking with other academic institutions. When my students do well and succeed, it not only makes me happy, but it’s good for our lab. I want to make sure you succeed!
Projects: When starting a new project, we’ll discuss projected timelines, milestones, and other expectations. Progress depends on many factors, including your course load and the scope of the project. We’ll continually be reviewing what you’ve completed, what progress should look like going forward, and if the work so far is acceptable (i.e. if you’ve been slacking off). I’m not a fan of what’s called “Least Publishable Unit,” which is producing work that is just enough to publish a paper in a low-tier conference or journal. You should aim for high-quality work that can achieve good impact in a top venue such as the Vis and CHI conferences. Sprinting to hit a deadline usually results in subpar work that’s unlikely to get accepted. While it’s important that you work hard and hit most of your targets, it shouldn’t come at the expense of good research. If you cannot hit a target, we’ll discuss it over and probably decide to push to the next reasonable venue.
Lab Time: I generally prefer students work in the lab, since that facilitates a more-focused work environment, promotes lab unity, and allows for more interaction and discussion together. That said, if you are making sufficient progress on your work, I’m rather lax on when you show up and leave (I usually don’t arrive before 10 am) and you don’t have to come in every day. In addition to 1-on-1 meetings to discuss your current projects and other academic obligations, group lab meetings will occur weekly. We use that time to review the prior week, update each other’s projects and provide feedback, give practice talks, etc. We also have occasional outside activities like lab dinners. At ASU, we do not tolerate any form of discrimination, harassment, or otherwise hostile conduct. Such actions are a good way to get expelled from the program.
Paper Writing: The bullet point about mentoring also applies to writing research papers. Technical writing like you see in research papers is difficult, and it can be doubly hard if English is not your primary language. Especially for your first couple of papers, you’ll start by producing a bulleted outline of all the critical points to make in the full paper. Once this outline is approved, you’ll create a rough draft well before the target venue’s submission deadline. The target timeline for a first draft is at least 1 month prior to the deadline. While this might seem like an enormous gap, you’ll be surprised at how long it takes to refine a rough draft into a polished and high-quality submission! We’ll go over your draft together and discuss edits for the second draft. You will probably have to make multiple iterations to improve the quality. At some point, your lab mates will also read the paper and provide feedback. This is a helpful step, because the more eyes that see it, the more issues get noticed and the better the overall product is. Best case scenario, everything is ready to go in plenty of time; the days leading up to the deadline are for doing minor tweaks and creating a demo video.
Misc. Lab Practices
Programming Languages and Software Stacks: While I have no hard and fast rules about programming languages and software stacks are required, normally we use popular libraries like D3.js, OpenGL, WebGL, and DeckGL for visualization, paired with layers like Flask, Node.js, SQL, etc. For projects that don’t require VR, high-end rendering, or specialized displays (such as tablet displays), web-based front-ends are generally easier, faster to develop, and more accessible compared to desktop applications, plus you can hook them into any programming paradigm you want via the backend server to do data analysis and management. If you’re new to full-stack dev work, that’s okay! You’ll pick it up as you go, though it certainly helps to have prior experience with it (especially front-end and/or interaction design).
Version control: We have a lab GitHub organization that is available for project management, or you may use your own repos. You should always store your code and data using some kind of version control, be it GitHub, BitBucket, etc.
Paper Writing: We write papers using LaTeX (unless there’s some necessary mitigating circumstance). Don’t worry if you’re not familiar with it; I wasn’t either until I began grad school! Overleaf is an online tool for LaTeX which makes it easy to share and collaborate on paper writing. For creating figures in papers (such as for user study results), we normally use libraries like R’s ggplot2.
Bibliography Management: Especially early on, you will read A LOT of research papers to learn what all is out there in the Data Vis field—past work, current trends, and future directions. I highly recommend you have some sort of log, journal, diary, or other management system for organizing the papers you have read. Ideally, you want something that allows for relatively fast review when you might later need to cite a specific paper and how it applies to your current work. When I was a graduate student, I kept a huge Google Doc where I would paste screenshots from papers and jot down summary notes and relevant quotes. Currently, I use a free software called Zotero, which can export bibtex files for LaTex papers.