Research

Cecilia Aragon's Research Group Archives

The following research group descriptions are archived because they are no longer offered, the faculty member is on sabbatical, or the group is taking a break. Please contact the faculty member or an advisor to learn more about these groups.


Data Science Ethnography

Big data, the data deluge, the information explosion... there have been many names to describe the overwhelming amount of data that is being generated in just about every scientific domain today. Data science is the term that has emerged to describe the study of the extraction of knowledge from this flood of data, and it can include elements of various fields from computer science to applied mathematics to human centered design and engineering.

However, little is known about the culture and human processes surrounding the emerging practice of data science. A recent five-year, $37.8 million award to UW, UC Berkeley, and NYU from the Gordon and Betty Moore Foundation and the Alfred P. Sloan Foundation seeks to address this gap.

In this research group, we will utilize ethnographic practices including contextual inquiry, interviews, and participant observation to delve more deeply into the culture of data science on the UW campus. We will participate in scientific efforts in astronomy, oceanography, sociology, and other exciting data-driven fields on the cutting edge of science today.

We are looking for students with a background or interest in ethnography, who are each interested in between 2 and 5 credit hours of credit/no credit grade of HCDE 496/596. This group will meet Mondays from 3:30–4:30 p.m. in Sieg Hall, room 420.


Distributed Mentoring in Online Fanfiction Communities

Are you both a fan and a hacker? Are you interested in studying how people learn from online fandom?

This ongoing research project studies informal learning in online fanfiction communities. We are looking for a small number of experienced programmers interested in fandom to join an existing research group. We have already published one paper on our research (arXiv:1510.01425v2) and are in the process of submitting others.

We suspect that the novel concept of distributed mentoring plays a positive role in fanfiction authors’ development as writers, and this quarter’s project attempts to quantify this effect. We intend to scrape stories, reviews, and associated metadata from fanfiction sites and apply quantitative techniques (machine learning, statistical analysis, data visualization) to investigate the relationship between distributed mentoring and writing quality (e.g., grammar, reading level). Applicants must have spent substantial time outside of class writing scripts to scrape the web and process text, in languages such as python, perl, or bash. No experience in machine learning or visualization is required, although it is a plus.


Qualitative Coding and Analysis of Affect (Emotion) in Text

Co-directed by Cecilia Aragon & Taylor Scott

We are studying creative collaboration in a distributed team of astrophysicists and have collected a large amount of longitudinal data in the form of chat logs. We have been qualitatively analyzing this data to detect and classify emotional content, relate it to events occurring in the group's history, and form a theoretical framework of Distributed Affect. Our initial methods have been successful and promising, and we plan to refine and verify them through further qualitative coding and analysis of the data.

We strongly encourage interested undergraduates to join this group, even if you have little or no experience with qualitative research. This is an excellent opportunity to be introduced to various methods of analyzing text data, and gain insight into the way that such research is carried out.

Participation in this research group should be a good opportunity to:

  • Gain valuable practice in qualitative coding of chat log data
  • Learn more about the application of methods and theoretical perspectives in qualitative data analysis
  • Apply visual analysis as a means of exploring a large data set
  • Discover how these methods can be applied to your own areas of interest and research

Visualization of Large Text Data Sets

The amount of informal text communication (e.g. chat, texting, microblogs) in the world is increasing exponentially. Submerged within this text data deluge lies a wealth of information that is potentially valuable to businesses, governments, social scientists, and all human communities. In this research group, we will develop text visualizations with a specific focus on visual concordances that can be applied to very large text data sets.

This will be a two-quarter directed research group with the goal of submitting a paper to Vis 2016 in March 2016. During the first quarter, we will sketch, stretch our visual imagination with hands-on design exercises and critiques, and build and test visualization prototypes in javascript and d3. During the second quarter, we will iterate on the research questions, refine our visual prototypes, conduct usability tests of our designs, and write a paper on our results.
 


Analyzing Online Community Data

We have been studying distributed creative collaboration in an online community of children creating programmable media such as games, interactive stories, music and art on a YouTube-like website developed at MIT (scratch.mit.edu). We are interested in analyzing chat log data from the site in order to develop a measure of learning effectiveness in such distributed communities. We will base our analysis on a theoretical framework developed by Turns and an already completed coding taxonomy of a subset of the data developed by Aragon.
Participation in this research group should be a good opportunity to:
  • Experience how theory is used to guide analysis of data
  • Understand how collaborative analysis of data can be organized
  • Learn a new set of theories (externalization of knowledge, creative resonance)
  • Learn about publication venues
Gain insight into what students are thinking about when they engage in educational activities
We are looking for a relatively small group of people who are each
interested in between 2 and 5 credits. The actual organization of the work will be based on the number of people interested. 
 

Visualization of Large Text Data Sets

The amount of informal text communication (e.g. chat, texting, microblogs) in the world is increasing exponentially. Submerged within this text data deluge lies a wealth of information that is potentially valuable to businesses, governments, social scientists, and all human communities. In this research group, we will develop text visualizations with a specific focus on visual concordances that can be applied to very large text data sets.

This will be a two-quarter directed research group with the goal of submitting a paper to Vis 2016 in March 2016. During the first quarter, we will sketch, stretch our visual imagination with hands-on design exercises and critiques, and build and test visualization prototypes in javascript and d3. During the second quarter, we will iterate on the research questions, refine our visual prototypes, conduct usability tests of our designs, and write a paper on our results.

 


Understanding and Analyzing Eye Tracking Data

The belief that eyes are the windows to the soul has driven the study of human gaze tracking for over a century. However, it is only recently that technological advances in eye tracking technology have led to dramatic reductions in both the intrusiveness and cost of eye tracking systems. This has led to a recent surge in interest in this technology, which is now widely used in such diverse areas as market research, usability, reading diagnostics, neuroscience, psychology, robotics, games, and more. Even a passing familiarity with eye tracking technology is a highly desirable skill in today’s competitive job market.
 
This research group will allow you to develop knowledge and familiarity with the theory and history of eye tracking, the process of conducting eye tracking experiments, the use of eye tracking equipment, and the analysis of eye tracking data. The group will produce a set of research posters based on their data analysis with the aim of eventually publishing their work in appropriate conferences.
 
I am looking for a relatively small group of people who are each interested in between 2 and 5 credits. The actual organization of the work will be based on the number and background of people who sign up. If you are interested, please send an email to Cecilia Aragon (aragon@uw.edu) with a few paragraphs describing why you are interested in the project, your background in eye tracking or other relevant research, if any, and the number of credits you are seeking. This group will meet Wednesdays from 2 - 3:30 in Sieg 427.

Reading Group: Remixing User Research Methods
 
Winter 2014
 
This reading group will explore user research that adapts and combines methods from both within and beyond HCI. By reading and discussing a broad range of studies demonstrating appropriate, contextualized user research design, students will expand their toolkit of user research approaches. Topics will cover the spectrum from formative user research of values and process to evaluation of technological interventions, but with an emphasis on open-ended, exploratory, and/or qualitative methods.
Facilitated by Katie Kuksenok, Computer Science PhD student (supervised by Cecilia Aragon).
 

Informal Learning in Online Fan Communities
 
The internet has opened up unprecedented opportunities for people of all ages to discover and connect with others who share their interests. Among the most popular interest-based communities are those that bring together fans of various media texts, including movies, TV shows, music bands, novels, and video games. Whether formed around classics like Star Trek, Doctor Who, or Blade Runner, or newer media texts such as Breaking Bad, the Twilight series, or World of Warcraft, these online fan communities make it easier than ever before for people to meet other fans and engage in discussions and creative endeavors around their mutual interests.
 
Though scholars have begun to explore the learning that takes place in online fandoms, we still lack a complete understanding of the skills youth develop through their fan-based activities; the roles that identity, motivation, and emotion play in young people’s informal learning online; and the novice to expert trajectories made available in different online fan communities. This research group will shed light on each of these areas of inquiry through an ethnographic investigation of online fan communities currently popular among U.S. teens. The group will produce a technical report of this investigation with the aim of publishing the work in appropriate conference proceedings.
 
We encourage both graduate and undergraduate students to join this group. Qualitative research experience (e.g., contextual analysis, ethnography) is desirable but not required.
 

Analyzing Emotional Content of Text-Based Communication
 
The automated detection and classification of emotion in text-based communication is an open problem, yet recent research has indicated the importance of emotion in collaborative creativity. We have been studying creative collaboration in a distributed team of researchers from an international astrophysics group studying supernovae to learn more about the expansion history of the universe. As a result, we have collected a large amount of longitudinal data in the form of chat logs. We have been exploring methods to analyze this data to detect and classify emotional content, and relate it to events occurring in the group's history.
 
Analysis of the data has been carried out through both manual and automated coding of the chat logs. The coding scheme we have developed is grounded in this data set, and informed by existing taxonomies of emotion. The group had developed an automated approach to this type of classification, which is now freely available and open-source, that uses the manually coded data as the training set for our classification algorithms. We will continue to build a corpus of coded data, and will be leveraging human interaction to train better automatic emotion classifiers via a ‘human-in-the-loop’ model.
 
Research conducted by this group has been published in multiple papers in competitive conferences, and all contributing group members are co-authors.
We strongly encourage interested undergraduates to join this group, even if you have little or no experience with this type of research. This is an excellent opportunity to be introduced to the various methods we are using, as well as a chance to gain valuable insight into the way that research is carried out.
 

Games for Good: Designing and Building Collaborative Games for Engineering and Science
 
This research group will explore the use of computer gaming for collaborative science and engineering learning and discovery. Students will be highly hands-on in developing and designing aspects of a game, using the Unity3D game engine to build new levels and environments, constructing narratives, and integrating player feedback as part of an iterative design processes. We are currently designing a futuristic crime scene investigation game that embeds bioinformatics concepts as players work together to save the world from a fictional bioterrorist organization.
 
Research questions we will explore include:
  • What makes for compelling game mechanics and narrative storytelling?
  • What is the role of social game play and how can game environments support collaboration?
  • How are affect and emotion supported in a game environment to promote greater scientific creativity?
Programming experience or graphic design experience is highly recommended (yet not required if the student is motivated to get up to speed quickly in one of these areas).
 
 

Cecilia Aragon
Social Media Qualitative Analysis: Methods and Practice (SMQA DRG)

Large data sets that reflect activity on Facebook, Twitter, and other social platforms can be vital for investigating human phenomena as well as using social media activity to understand product or policy impact. However, existing HCI research methodologies do not scale in a straightforward way, resulting in ad-hoc data collection and analysis that may or may not achieve desired goals. This research group will develop an integrative methodological framework for analyzing social media data, focusing on text in context of time, network structure, and other associated metadata. In this group, we will have two complementary foci:
  • Methods
    • developing an integrated methodological framework that can be used by researchers in designing and evaluating studies of social media texts
    • relating existing practice with epistemological bases for existing empirical social science methodologies
    • involves understanding related methodological research in HCI and extensive analysis of existing literature on social media data studies
  • Practice
    • identifying and engaging with stakeholders in social media data analysis (including both researchers and, for example, managers who make decisions on the basis of research findings) across both academia and industry
    • contextual inquiry into the practices of researchers and their tool ecosystem
    • involves understanding related CSCW studies of similar populations, designing and distributing surveys, conducting interviews, and observation
In the course of this year-long project, we will spend the fall digging into the methods branch and designing studies for the practice branch, and conduct those studies during the winter and spring. This work will culminate in CSCW and/or CHI submission(s) and any other external deliverables that we determine valuable to relevant stakeholders.
HCDE professional masters and undergraduate students are especially encouraged to apply. Meeting time and location will be determined based on the schedules of group members. 
 

Understanding and Analyzing Eye Tracking Data
 
The belief that eyes are the windows to the soul has driven the study of human gaze tracking for over a century. However, it is only recently that technological advances in eye tracking technology have led to dramatic reductions in both the intrusiveness and cost of eye tracking systems. This has led to a recent surge in interest in this technology, which is now widely used in such diverse areas as market research, usability, reading diagnostics, neuroscience, psychology, robotics, games, and more. Even a passing familiarity with eye tracking technology is a highly desirable skill in today's competitive job market.
 
This research group will allow you to develop knowledge and familiarity with the theory and history of eye tracking, the process of conducting eye tracking experiments, the use of eye tracking equipment, and the analysis of eye tracking data. The group will produce a set of research posters based on their data analysis with the aim of eventually publishing their work in appropriate conferences.