Clemson Research Experiences for Undergraduates

About Karina

Hi, I'm Karina! I was born in Pickerington, Ohio and raised in Concord, North Carolina (right outside of Charlotte). I am a junior majoring in Industrial Engineering at Clemson University. I am Cuban, Colombian, and Finnish and very proud of my heritage. My first name is Finnish, but it is also a common hispanic name, which my parents didn't know until many years after naming me. Coincidence! I am an avid sports fan; soccer is my favorite to play, and football is my favorite to watch. You can find me at any home game in Death Valley! Aside from all Clemson athletics, I cheer for the Pittsburgh Steelers, Carolina Panthers, and the Pittsburgh Penguins. I also love to dance! I used to dance competitively, and I now spend some of my free time participating in Clemson Dancers. We put on an amazing showcase every year, and I'm super excited to be Vice President of the club next year! I am part of the Calhoun Honors College here at Clemson, and I am very active in the honors college community service group. This past year, I was named the Janine Anthony Bowen '89 Undergraduate Scholar, an award assigned by my department. I have also received the Dwight D. Gardner Scholarship, awarded by the Institute for Industrial Engineers, for the 2015-2016 school year. When I'm not working, learning, or being active, I love to spend time with my 10-month-old puppy, Echo! He loves people more than any other human or dog I have ever met, but he will bolt to his favorite hiding spot any time he hears a loud noise. Owning a puppy has been one of the most challenging yet rewarding parts of my life. Your parents ARE NOT kidding when they tell you that owning a dog (or almost any animal) is a huge responsibility! I'm really looking forward to my research this summer, as I have participated in a few research experiences before but never directly in my field of study. I can't wait to learn and grow in another aspect of my life!

Project Description

Using Visualization to Effectively Capture Multimodal Data Over Time

This summer I will be using visualization techniques to figure out a standard for presenting multimodal interfaces on paper (for example, a research paper). Research of multimodal interfaces is becoming increasingly popular, but the complexity of such interfaces can be difficult to present in a two-dimensional format. By developing a type of "standard," researchers in the future will have an easier time presenting their work.

Week 1

This week was mostly consumed with learning how to develop a webpage and how to use Linux and Palmetto. On Wednesday I began literature review for my project. In order to understand what I will be working on, I need to gain some base knowledge in the field of multimodal interfaces. On Thursday I was able to meet with my mentor, and we discussed expectations and project basics. She also gave me a few papers to read through to continue my literature review. For the rest of the week, I read through the provided papers, and began to use Google Scholar and Clemson library databases to search for more studies and articles in multimodal interfaces relating to human factors.

Week 2

This week began with exactly what I had been waiting for: a lecture on visualization. I was awaiting this lecture so that I would have a better understanding of what was expected of me this summer. Throughout the rest of the week, I attended research and enrinchment lectures while completing even more literature reviews. We also had a webinar on graduate school! To close out the week we went on a tour of the Digital Production Lab, which was a pretty cool part of Clemson that I have never seen before. Though I haven't gotten into the thick of my work yet, I suspect that will change next week, so stay tuned for more information about the development of my project!

Week 3

This week started out a bit slow for me, but picked up pace towards the end! My car broke down in my hometown over the weekend, so I had to purchase a new car and missed work and sessions on Monday. On Tuesday, I began catching up. Dr. Riggs and I had multiple meetings this week so we could get on the same page and begin looking through the data. We decided to use data from an airline pilot study that she helped conduct in 2012. I'm starting by looking only at the eye tracking data, and I was given free-reign to visualize that data. I decided to try Gephi, but since I have never used it before, I had a bit of a learning curve. I spent a lot of time using one data file and adjusting different variables and parameters in Gephi to see the capabilities. In the past, the eye tracking data was plotted using a program written in MATLAB by one of Dr. Riggs' colleagues at the University of Michigan. When I finally got my data plotted in Gephi, it matched up (aside from being upside down). I then started working on adjusting the size and color of the nodes to portray different variables like duration of the "stare." Finally, I began working on Dynamic Topography in Gephi so that I can potentially show the data over time. For example, nodes will appear and disappear to simulate the participant's eye movements during the study. I still have a lot to learn!

Week 4

This week was very busy! It began with a class about developing an elevator pitch, and afterward, I continued to work in Gephi to figure out how to make the data dynamic. I did this by using the "Dynamic Graph Example," and I can now add nodes over time; however, the past nodes do not fade with time. Next I began adding event placeholders in each of the data sets so that when I begin to add the auditory and tactile events, it is much easier to decipher. I also fully graphed five of the data sets to give Dr. Riggs a better look at what I was doing. Upon further inspection, I realized that both the X and Y coordinates were printing into Gephi backwards, so next week I will fix that issue. I also worked extensively on my abstract and midterm PowerPoint slides this week. I believe that I have a very good product! Next week I will finish graphing each of the data sets then move on to graphing each individual event.

Week 5

This week was rather stressful with the impending midterm presentations and subsequent conference attendee selections. I practiced my presentation multiple times alone and in the presnce of my mentor, and continuously adjusted according to critiques and suggestions. I felt that my midterm presentation went beautifully, and I was happy that I was able to field questions appropriately. Below is an example of my data that shows what I have accomplished so far. The data has been mapped, and the nodes are sized according to stare duration. The color of the node darkens over time, and they also appear over time using the dynamic topography mode. (I apologize for the watermark; I do not currently have another way to collect these videos from my screen). After the presentation, I began pushing foward to figure out how to incorporate the different modalities. Firstly, I corrected the x and y coordinate mapping issue, and I began to create a pseudo-screen overlay, which is mostly guess and check work. I also started to determine what would be best to represent each modality (maybe node color) and each event type (maybe node color). Node shape would be an ideal variable; however, Gephi does not seem to be capable of this yet. My tasks for the rest of the summer include: representing the modalities and event types, grouping by event and flight types, and creating a process for representing multimodal data for future use.

Week 6

This week was a bit shorter because of the observance of the 4th of July, but there was still a lot of work to get done! I took a lot of time to perfect the node coloring according to the modality being used, and I also colored the edges according to the event type (comm, traffic, or alert). This is A LOT of work for only one flight for one participant! I also continued to work on the screen overlay. After hitting many speed bumps, I think I have found a process that will give me a good estimate for the screen positions. While it will not be exact, future researchers may keep their displays in mind so that they can measure their locations in reference to the eye-tracker coordinates (or other measurement pertaining to their study). I hope to have the overlay completed by the beginning of next week so that I can begin my next steps. At the REU mini-conference in Charleston this weekend, I had a great experience and learned a lot. After talking to many of the participants, I discovered that it may be advantageous to create a function to input my data into Excel instead of doing it manually. I also learned about the ErgoEasy app they are working on at Auburn, which piqued my interest because ergonomics is a huge part of industrial engineering. I look forward to hearing about that app in the future! Additionally, I learned that talking for 45 minutes straight is SO hard!!! You can see my slides from my lightening talk below.

Week 7

This week I made some very big strides! I finally finished the screen overlay for my data, so I was excited about that. I finalized the node and edge coloring for the data...then I stared at the other 13 files I had to perform all of these time-consuming actions on. Instead of spending unnecessary time on that, I decided to write a MATLAB program to do it all for me!!! It took me around 8 hours to write, but it would have taken way longer for me to do each of the files individually. The program basically took the original data file, and created two files (a node and an edge file) with TWO clicks of a button! Then all I had to do was save these files as CSV for input into Gephi! I was very proud of my program; I hadn't touched MATLAB in over a year, so I had to work to remember a lot then learn even more. Thank goodness I kept all of my notes from that class! After obtaining all of the new files, I collected multiple images for each flight (overall, events only, and each event individually). I also started collecting videos of the data evolving over time; I will still have to finish that this weekend or early next week. Finally, I began work on my final poster! Lots of work to get done before FINAL PRESENTATIONS next week.

Week 8

The final week was so bitter sweet! I started the week with completing the rest of the data animation videos. I also worked very hard to complete my poster, and I think it looks great! I got some feedback from many sources that helped me improve it as well. I spent the rest of the week working on my final paper and final presentation. I have finished my first draft of my paper, so over the next few weeks I will be editing it with the help of my mentor until it is due on August third. My final presentation went great! I handled my questions well and now I cannot believe that my eight weeks with all of these amazing people is over! I was excited after the final presentation when my mentor forwarded me an email from Janet's mentor Dr. Qiushi Chen asking if I could provide my Matlab code so that he could look it over and create something similar for his data! I'll be sending that out to him, but I have to make sure to comment the code first!!! Anyway, I am going to miss everyone that was a part of my experience this summer, but I hope to see them doing great things in the future.

Last updated: 7/24/2015