Past REU Program Projects

Summer 2015 Project Descriptions

PeerPresents

Our team seeks to improve peer feedback on student presentations. Many college courses require students to give in-class presentations. This raises a number of challenges. Students who watch the presentation are typically not participating actively. Professors can be overwhelmed trying to provide feedback while also managing the class. The presenting students would benefit from receiving more feedback than the professor can provide. This project introduces PeerPresents, an in-class peer feedback system we developed to improve the feedback process for in-class presentations. We are looking for undergraduates to help develop the tool, implement classroom studies, and analyze data. Ideal candidates will be familiar with Javascript, Node.js, CSS, MongoDB. You will have the opportunity to work with experts in HCI, learning science, and game design and to creatively influence the direction of the research!

People: Jess Hammer, Steven Dow, and Amy Shannon

Typing at Natural Speaking Rates on Existing Keyboards

A vital accommodation for people with deaf and hard of hearing people is real-time captioning. The only reliable approach for this currently is to employ professional stenographers who train extensively to type at more than 250 words per minute using special chorded keyboards. This goal of this project is to develop approaches for training people to become real-time stenographers on existing keyboards. This will involve work across human-computer interaction, including technical innovation, game design, language technologies, and innovation in interaction techniques.

We seek students with interest and experience in one or more of:

  • web programming (primarily front-end)
  • user interface software and technology
  • human-computer interaction
  • accessible technology

People: Jeff Bigham

Crowdsourcing Accessible Research Papers

Nearly every research paper that is written is made available in PDF format. Unfortunately, most of these PDFs are created in a way that people will disabilities cannot easily use. The goal of this project is to combine automated technology and crowdsourcing to make PDFs accessible so that everyone can benefit and learn from them. This project has the potential for huge near-term impact.

We seek students with interest and experience in one or more of:

  • strong programming skills (C, Java or Python)
  • familiarity with low-level document formats
  • Latex or Postscript
  • accessible technology

People: Jeff Bigham

Crowdsourcing Solutions to Hard Web Accessibility Problems

The Web offers a great opportunity for people with disabilities, but a lot of web content is not created in a way that people with disabilities can easily use. We are interested in how crowdsourcing can be used to solve difficult problems the web accessibility space, such as how simplifying text on the Web for people with cognitive disabilities, improving usability for people with motor impairments, or describing complex images and figures. This project involves combining automated approaches with crowdsourcing. The goal is to produce and release web-based tools that solve some problems.

We seek students with interest and experience in one or more of:

  • web programming (front-end and back-end)
  • human-computer interaction
  • crowdsourcing
  • language technologies and computer vision
  • accessible technology

People: Jeff Bigham

Urban Analytics

What can millions of pieces of geotagged social media tell us about a city? We’re interested in going lots of analyses and creating visualizations to help understand people’s behaviors in cities. Examples include business analytics (Who comes to my store? How far do they come?), inferring commute times, and improving reverse geotagging algorithms (given a GPS location, what venue are they at?).

People: Jason Hong

Social Computing for Design Critique

In this proposal, we investigate how online technology can support design critique and enhance student learning in design studio courses. We propose to create a repository of online tools that support design critique. The repository of online critique tools, the content of the critiques, an assessment of the critique comments, and the quality of the design projects create a valuable corpus of data to study design critique processes and improve online critique tools. The content of the design critiques will be used to study the design critique process and extract design critique heuristics. The critique heuristics of experts can in turn be used to scaffold novices to learn to conduct better critiques. The repository will also serves as a platform for future research. We are looking for undergraduate students with good communication and design skills to help create and evaluate such an online repository.

People: Steven Dow, Jodi Forlizzi, Peter Scupelli

Visualizing social interactions
 

We’re designing and building novel info viz of social interaction using sensor data (Kinect and physiological sensors). For example, we are building “rapport-meter,” an application that senses gestures and body position of two people having a F2F conversation. We want to extend it to give feedback about how well the interaction is going. We want to explore what happens when you use the ‘rapport meter’ in real-life situations such as job interviews, physician consultations or on first dates. The project will involve creativity AND programming. Familiarity with processing, javascript, flex, flash, interaction design skills, love of data, and interest in social interaction and/or physiological sensing would a plus.

People: Min Kyung Lee, Laura Dabbish

 

Measuring emotional response to online interactions

Our project examines the deep structures of collaboration – the physiological signals associated with different forms of technological support and technologically mediated interaction. We are using galvanic skin conductance, heart rate variability and camera based face tracking to measure emotional response to social information and online interactions. The REU students on this project will learn how to collect physiological data on emotional response and participate in designing and running experiments to measure emotional response during online interactions.

People: Min Kyung Lee, Laura Dabbish

 

Digital Footprints

The Internet increasingly reveals personal information about people to each other and to third parties. The goal of this research is to understand how people manage their digital footprints on the Internet, and explore different interaction techniques to help people manage their information. You will learn about HCI and social science research, be involved in designing and running user studies, analyzing data, and potentially co-authoring a conference paper.

People: Ruogu Kang, Laura Dabbish, Sara Kiesler

 

Social Transparency and Critique

We’re designing a series of studies to examine how people respond to critique, comments, and feedback on their work depending on the amount of contextual information available about the person critiquing them. We will use several physiological indicators such as GSR (Galvanic skin responses) and ECG (electrocardiogram), facial and bodily expressions to measure such responses. The REU students on this project will learn how to conduct lab experiments with human subjects. They also learn to carry out physiological measurements including operating the equipment and analyzing the data.

People: Mary Nguyen, Laura Dabbish, Steven Dow

 

Technology Enabled Role Play

College professors frequently use role-playing in class as a way to engage their students and enhance the learning experience. However, role-playing often creates an unbalanced classroom, where the majority of students are observing and only a few actors are “engaged”. This project investigates using technology to engage all students during classroom role-play. REU students can expect to help design and build a prototype technology and run user studies. We are seeking students with coding experience and good writing/communication skills. You will also have the opportunity to work with experts in HCI, learning science, and game design and to creatively influence the direction of the research.

People: Steven Dow, Jess Hammer, Amy Shannon

http://protolab.cs.cmu.edu/Projects

 

Do It Yoursef for Other People

Millions of people with disabilities need assistive products and services to help them live their lives. However, traditional production models and economies of scale fail to produce the kinds of personalized and affordable solutions people need. This project investigates how we can motivate DIY makers to volunteer their skills to create bespoke assistive technology devices and services. The research also works towards a vision for a peer-to-peer service economy where people with specialized skills collaborate on virtual teams to work on well-scoped challenges. We are looking for undergraduate students with good communication skills to help analyze and synthesize field research and with HCI/CS skills to design and build an interactive system to guide maker teams through design processes.

People: Steven Dow, Jodi Forlizzi, Eunki Chung

 

Comparing Approaches for Assessing Creative Work

We are interested in developing scalable techniques for assessing the creativity of digital designs (e.g., wireframes, storyboards, or poster designs). These techniques might be used in design classes, MOOCs, or for conducting controlled research experiments. We currently have a number of different techniques that we are interested in comparing and are looking for an undergraduate or masters student to help test and identify the best methods. This research will consist of launching studies on Mechanical Turk and analyzing the data. Ideally, participants will have knowledge of how to create web pages using HTML and basic knowledge of programming (for data manipulation). You will gain experience running studies on Mechanical Turk, conducting data analysis, and writing a research paper. You will also have the opportunity to work with expert HCI researchers and to creatively influence the direction of the research! Former student collaborators have gone on to top graduate programs, and have co-authored full research papers accepted to top-tier HCI conferences. If interested, please email your resume and a short description of your experience with web development and research experiments.

People: Chris MacLellan, Steven Dow

 

Redesigning Online Health Support Communities

A high percentage of people with chronic or life-threatening diseases participate in online health support communities to exchange informational and emotional support with other patients and caregivers. We are working with the American Cancer Society to redesign their Cancer Support Network to make it more effective. The Cancer Support Networks is the largest online support community in the world for people suffering from cancer. Do you want to improve cancer survivor’s quality of life by applying your skills in interaction design, usability analysis and testing, social science research, machine learning or programming? Our goal is to improve the experiences people have when participating in online health support groups. We are seeking undergraduate students as research assistants, Students with majors in human-computer interaction, psychology, design and computer science or related areas are welcome.

People: Robert Kraut

 

Socializing newcomers to online communities

The majority of new members quickly leave online communities, be they health support groups, gaming communities or the editors of Wikipedia. This project is analyzing data from Wikipedia to understand the conditions that keep new editors contributing. We are also conducting experiments within Wikipedia to increase retention. This summer we are focusing on how joining a community with a cohort of other newcomers influences retention and success.

People: Robert Kraut

 

Social Ideation

This project investigates how social computing systems can better support real-time creative collaboration, whether in teams or with crowds. There are a number of exciting opportunities to design, build, and test social computing approaches to improving creativity (e.g., supporting group awareness, motivating participation through gamification, support for real-time collaborative synthesis of ideas) within our current prototype system. Students can expect to be heavily involved in all phases of the design process, from conceptualization to iterative development and testing of system features (both in-person testing and A/B experiments online). There is also the potential for co-authorship on full research papers for top-tier HCI conferences! The ideal candidate will have interests in social computing and/or creativity, and prior experience with interactive web application development (in particular, Node.js and/or Meteor.js) and/or popular visualization frameworks (e.g., D3.js).

People: Joel Chan, Steven Dow

 

Distributed creativity

Innovation in is often driven by analogy. The opportunities for finding fruitful analogies are exploding with the availability of repositories of ideas ranging from scientific papers to product ideas to the text and video resources. For example, in 2005 a car mechanic invented a heralded technique for saving babies stuck in the birth canal that he adapted from a method for extracting a lost cork from a wine bottle seen on a YouTube video. However, our ability to find and use analogies is severely hampered by people’s cognitive limits. Instead of relying on a single individual to find and apply an analogy, our research is examining how to distribute analogical processing across multiple individuals. This summer we will be working with the product design firm Quirky to understand how best to destribute analogical design processes across multiple people. These processes include representing problems and potential solutions in an abstract way to facilitate comparison, identifying analogies in domains distant from an initial problem and applying ideas from these analogies to solve the original problem. You will help design, implement and analyse experiments to test the effects of different ways of distributing and recombining these processes across people.

People: Lisa Yu, Niki Kittur, Robert Kraut

 

CrowdVerify – Using the Crowd to Simplify Privacy Policies and Terms of Use

The goal of this project is to investigate how to use crowds and voting-based systems to help simplify privacy policies and terms of use statements. Imagine that you split one of these policies into sentences, and then have people choose which sentence is more important. Over time, you can build a model of what people are likely to say is and is not important. It’s sort of like kittenwar.com, except for policies. We’re looking for two undergrads with mad computer science skills to help us on this work. One would help create a crawler to gather policies for the top 1000 sites, as well as create a web site that people can try to test our models. The other would help adapt some software and deploy a site where interested community members could vote on statements, helping us in gathering more data about people’s preferences.

People: Jason Hong

 

PrivacyGrade – Ensuring the Privacy of Smartphone Apps

PrivacyGrade.org is our site for grading the privacy of smartphone apps, and has been featured in the NYTimes, CNN, Forbes, BBC, and more. We’re looking for two undergrads to help advance the state of the art with respect to privacy. For the first project, the goal is to create machine learning models that can help predict whether an app seems to be targeting young children. This is an important step in helping to enforce the Children’s Online Privacy Protection Act (COPPA). This project will require lots of Java programming skills and some familiarity with statistics and/or machine learning. For the second project, the goal is to help evaluate people’s sensitivity to different kinds of data abstractions. For example, most people are probably ok with sharing with any app that they are at “home” or “work”, but probably not their exact location. Can we quantify these different concerns? This second project requires skill in setting up and running user studies.

People: Jason Hong

 

Social Cybersecurity

How do people learn about cybersecurity? What do they talk about regarding cybersecurity? What kinds of social techniques can we apply to influence people’s awareness, knowledge, and motivation to be secure? We’re interested in deploying a cybersecurity quiz with some subtle tweaks to influence people’s behaviors (e.g. showing information about what experts thought, how other people did, etc). We’re looking for one or two undergrads who can help improve our software, deploy it, popularize it, and do data analysis on the results.

People: Laura Dabbish, Jason Hong

 

Multitouch Visualization

DataSquid is a research project that reimagines data visualization for mobile multitouch devices, letting users sort, sift, and stack data with their fingers. We’re looking for one or two undergrads who can either 1) build use cases by scraping data from the web or APIs and then generate compelling usage scenarios; or 2) run user studies evaluating the effectiveness of the tool. For example, usage scenarios may include visualizing yelp, mint, airbnb, and many other real world data sources.

People: Jeff Rzeszotarski, Niki Kittur

 

Document Triage with Tabs

How many times have you looked at your browser and thought “Wow I have way too many tabs open”? Tabs are messy, sub-optimized way for individuals to organize activities and sub tasks, and often they have to resort to using tools like Microsoft Word or Evernote to help. We are interested in improving how individuals use browser tabs during information seeking by leveraging multiple visualizations, automatic categorization, and behavioral tracking. You will help design, implement, and evaluate browser tools that will monitor tab behavior and assist users with quickly categorizing and filtering information as they use the web.

People: Niki Kittur, Nathan Hahn, Joseph Chang

 

Visualizing Wikipedia History

Wikipedia is quickly losing users, in large part because new users are quickly overwhelmed with the massive histories of previous contributions to an article. If they don’t pay attention to these, their work is often rejected, and they leave. This project involves developing novel visualizations and tools for exploring the history of Wikipedia. The work will include developing web-based tools to visualize massive amounts of data, and deploying/testing tools with live Wikipedia data. Your efforts may have an impact in helping people to edit Wikipedia and other collective knowledge stores, as well as resulting in publications at top HCI conferences.

People: Jeff Rzeszotarski, Joseph Chang, Niki Kittur

 

Summer 2013 & 2014 Project Descriptions:

Project name: Digital Footprints
The Internet increasingly reveals personal information about people to each other and to third parties. The goal of this research is to understand how people manage their digital footprints on the Internet, and explore different interaction techniques to help people manage their information. You will learn about HCI and social science research, be involved in designing and running user studies, analyzing data, and potentially co-authoring a conference paper.

Desired skills/background:

  • Prior experience with interviews, questionnaires, or lab experiments
  • Statistics and data analysis skills
  • Strong interest in computer and Internet; web programming skills would be a plus
  • Good writing/communication skills

Project name: Social Transparency
Online social environments support a new form of digital transparency because of the richly detailed data they store about artifacts, activities on those artifacts, and people. Activities anywhere in the world can be just as visible as those undertaken locally.  Activities from the recent or distant past are as easy to “see” as those in the current moment. In this project you will help examine how to design social transparency to enable learning from the actions of others. We will investigate how the design of activity trace information influences the ability to learn from the behavior of others. Ideal candidates will have web programming experience and an interest in behavioral science research.  You will learn how to design and conduct research, get exposure to data analysis and participate in writing a publication summarizing the results of the study.

Project name: User Interaction on Facebook
Online social networks can introduce “context collapse”, where one’s different audiences are all present in one place. We seek two REU students to work on an experimental project assessing how users interact with homogenous vs. heterogeneous friend groups on Facebook and whether personality traits influence user interactions. The REU students will assist with participant recruitment, experimental and survey-based data collection, and quantitative and qualitative data analysis. Educational benefits include gaining experience with experimental and survey methodologies, developing quantitative and qualitative data analysis skills, and assisting with preparation of a conference paper.

The ideal background of these REU students would be a major in human-computer interaction, computer science, cognitive science, or psychology with prior experience running user studies, programming skills (e.g., Python or Java, JavaScript, HTML, CSS, PHP), and at least one course in statistics.

Project name: PrivacyGrade.org: Grading the Privacy of Smartphone Apps
Many smartphone apps collect much more data than you might expect. For example, Angry Birds collects location data, and there are flashlight apps that use your phone’s unique ID. PrivacyGrade.org is a web site we are working on to displaythe results of our analysis on the privacy of hundreds of thousands of smartphone apps. We’re looking for a talented undergrad who can help us improve the design of the web site, implement ways of getting crowd feedback, and/or run user studies to improve the design of the privacy information about each app.

Good skills to have include any subset of interaction design, HCI, user studies, stats, programming, and databases.

Project name: Collaborative Search and Visualization
The goal of this project is to fundamentally change the way that information is consumed online; instead of ³searching² for information our goal is to support ³synthesizing² information. You will be involved in building a novel web interface for collaboratively searching and visualizing online information. Your efforts could have an immediate real-world impact, changing the way that people make sense of information as well as resulting in publications in top HCI conferences.

  • Ruby (preferred), Python, or other server-side scripting languages
  • CSS3, LESS/SASS, DOM, modern web standards, cross-browser compatibly
  • Familiarity with source control systems like Git for coordinating development
    €  – User interface programming and design skills
    €  – Experience with SQL/database storage
    €  – Experience with optimizing and deploying apps a plus

Project name: Crowd Programming
You will be involved in building tools and running experiments on Amazon¹s Mechanical Turk. The work will include building an interface and backend to run experiments on Mechanical Turk, collect data, and extend automated tools to manage tasks. Your efforts could have an immediate real-world impact, changing the way that experiments and work are conducted as well as resulting in publications in top HCI conferences.

  • Must be proficient in Python, SQL, Java, and Web programming (e.g., PHP)
  • as evident from past projects and coursework
  • User interface programming and design skills highly desired

Project name: Social Cybersecurity Quiz Game
In this research we are using social influence tactics to design interventions that promote safer cybersecurity behavior. Your summer project would be to develop a fun and engaging interactive quiz game where people can learn about other people’s security behavior across a population of users. The goal is to design, prototype and deploy this quiz website during the summer based on our previous qualitative research and social influence theory. You’ll gain valuable experience designing an interactive system and fun online social experience, and build your knowledge in social psychology theory and findings. A student with some background in web programming and interest in cybersecurity, social psychology and game design would be most suited for this project.

Project name: Crowdsourced support for visual design critique and learning
We are seeking a motivated student to help us develop and evaluate CrowdCrit, an online system that helps crowd workers learn design knowledge and provide valuable feedback to designers. We have a basic infrastructure running for CrowdCrit so now is the perfect time to join us in adding exciting new features. Potential student projects include opening up CrowdCrit to large-scale public participation, adding collaboration support so crowd workers can critique together, and exploring new interfaces for helping crowds learn complex skills.

The ideal candidate will have programming experience, especially in a web environment. (CrowdCrit is built in Python, Django, and HTML5.) However, strong candidates with experience in UI design, visual design, or data analysis will also be considered. An interest in social computing, design, or creative technologies is highly recommended. The student will gain  valuable experience prototyping social computing systems, translating learning theory into software features, and leveraging crowdsourcing platforms (e.g. Amazon Mechanical Turk) to conduct behavioral research.

Project name: Multitouch Data Visualization
You will be involved with building visualization tools for mobile devices that use physics to help people understand data. For instance, instead of filtering data, users might physically separate points using their fingers. We hope to expand existing prototypes (check out youtu.be/mZYj87qkaqA) to include new ways of visualizing data, help people collaborate across multiple devices, and support new platforms. Your efforts may have an immediate impact on consumer data visualization tools, as well as resulting in publications at top HCI conferences.

  • Must be proficient in mobile platform development
  • Proficiency with Objective C is ideal
  • Design and UI development experience highly desired

Project name: Crowd History Visualization
While Wikipedia is a tremendous resource, its size is also a burden. Pages have such large collective histories that it is nearly impossible to understand what went into a page, and how one can contribute. You will be involved with developing novel visualizations and tools for exploring the history of Wikipedia. The work will include doing large scale machine learning, developing web-based tools to visualize massive amounts of data, and deploying/testing tools with live Wikipedia data. Your efforts may have an impact in helping people to edit Wikipedia and other collective knowledge stores, as well as resulting in publications at top HCI conferences.

  • Proficiency in Python, Ruby, or other server-side scripting languages
  • Javascript, CSS, HTML, cross-browser compatibility experience
  • Familiarity with developing and deploying web-based apps
  • UI programming/design skills a plus
  • ML/NLP or large scale systems experience desired

Project name: Analyzing Crowd Behaviour
You will help build tools that observe, interpret, and act on the behavior of massive crowds of workers doing lots of small tasks. For instance, can we recognize when workers are confused by poor instructions or when a worker lacks the skills necessary to complete a task? The work will include building tools/interfaces for observing crowd behavior and running experiments on Amazon’s Mechanical Turk. Your efforts could have an immediate real-world impact, changing the way that experiments and work are conducted as well as resulting in publications in top HCI conferences.

  • Proficiency in Python, Ruby, or other server-side scripting languages
  • Javascript, CSS, HTML, cross-browser compatibility experience
  • Familiarity with developing and deploying web-based apps
  • ML/NLP or large scale systems experience a plus

Project Name: Exploring crowd-driven innovation
Industry and government depend on higher education to teach professional skills and to prepare students for careers in innovation. Yet, current approaches to innovation education fall short in helping students see a connection between design choices and real-world outcomes. Our research explores a new model for innovation education where students interact with potential consumers and experts throughout a design process.

We have spots for two students on this project. One student will help analyze data collected during a recent course where ~60 students used various social media to interact with stakeholders throughout a design process.  The intern will learn techniques for qualitative and quantitative data analysis. Another student with web-design skills will work on designing, testing, and iterating on our collection of crowd-based classroom activities. Through an iterative approach, the intern will refine the activities to make them easy to deploy and track from our website. Together, they will work our team to understand how factors such as the anonymity and scale of online responses affect how students shaped their innovation concepts.

Project name: Socialize Newcomers to Wikipedia

Project description: Have you ever tried to join a new group, but you didn’t understand the culture? Help conduct psychology experiments to ease students into editing Wikipedia as part of a class. Experiments focus on socialization, social identity, and small group interactions.

Relevant undergrad major or skills:

  • Experience or desire to interact directly with participants as a confederate.
  • Or alternatively, web programming skills (ajax, javascript, php, sql), to help design and implement experiments.
  • Background in social psychology helpful but not required

Project name: Social Search

People spend an enormous amount of time searching online for information. But, we find not only are we using out-dated tools to find and share information, but most often than not, someone else has carried out the exact same search. Here at Carnegie Mellon University, we are developing the next generation of search systems that leverage other people’s efforts and summarize and present information in an easy to digest format.

As part of the internship you will be involved in the initial design and conceptualization of new search tools and methods, to implementation and testing, to running user studies, to analyzing data and writing research papers. We expect this work to have immediate impact both in academia and in the real world, resulting in publications in top HCI conferences.


Project name: Strategies for Crowdsourcing Complex Design Work (2 students)

One student with a design background will help to prototype and design materials for the critique system. This includes a hierarchical list of visual design principles synthesized from a review of design textbooks and articles. The student will also help prepare a set of example designs with intentional flaws, corresponding to specific design principles, which allow us to instruct novices how to identify a variety of critique types with precision. The student will assist with the creation of pre- and post-assessments that measure individual design expertise, and will help us determine the efficacy of our learning techniques and generalizability of our findings. In these roles, the student will gain valuable exposure to design research literature and study methods, particularly experiment design and evaluation metrics.

Another student will help develop the critique system and run empirical studies. This student will assist with building and testing a web-based system with three major components: (1) training modules where crowdworkers learn design principles, (2) a critique interface where the crowd provides feedback on submitted designs, and (3) a review interface where the submitter explores the aggregated crowd critiques. The student will gain valuable experience prototyping social computing systems, translating learning theory into software features, and leveraging crowdsourcing platforms to conduct behavioral research.

Both students will help recruit participants, including designers with various levels of expertise and groups of crowdworkers, and proctor lab studies and online experiments.


Project name: Exploring crowd-driven innovation education (2 students)

One design-oriented student will work on designing, testing, and iterating on our collection of crowd-based classroom activities. Through an iterative approach, the intern will refine the activities to make them easy to deploy and use in a variety of classroom settings. The student intern will gain valuable exposure to new crowd-based techniques for innovation and to traditional HCI research methods.

Another student with computer science skills will help design a study and a web-based system to track the use of our crowd-based activities. We will deploy our activities and recruit instructors to try these activities with their own students.  The research intern will develop a study plan and a technical infrastructure for gathering input from instructors and students who download our materials.  The student intern will get experience with study design, web development, participant recruitment, and data analysis.


Project name: Idea Reuse and Crowd Creativity

The project is about crowd creativity. We are studying the process by which inventors generate and develop ideas online and ways of making these ideas useful to other inventors.

As part of the research team you would be working with HCI researchers in the Social Computing Lab to collect data and conduct experiments. The ideal candidate should have strong programming skills including Python, PHP, Javascript, CSS, and HTML. Preferably the candidate has interest in collecting and analyzing qualitative data and running experiments too.


Project name: Information sharing for collaborative analysis

Skills/background desired:

  • Prior experience conducting lab experiments (but we will teach you, if you don’t have this experience)
  • IRB investigator credential (you will need to do the tutorial if you haven’t passed the test–it takes a couple of hours)
  • Computer skills – modest, to adapt the web interface to our studies if needed
  • Data analysis skills, statistics are helpful
  • Good writing/communication skills

This project is looking at how groups of investigators (we use the example of detectives) can share information and hypotheses to reach better and quicker solutions. We have found in previous research that a detective who shares bad information with other detectives can harm their analysis. We are doing studies to find out how we can improve analysis and particularly what tools would help analysts predict or discover early that the information they are getting is bad.


Project name: Anonymity on the Internet

Skills/background desired:

  • Prior experience conducting online surveys desirable, e.g., using MTurk or Survey Monkey.
  • IRB investigator credential (you will need to do the tutorial if you haven’t passed the test–it takes a couple of hours)
  • Computer skills – modest, to adapt the web interface to our studies if needed
  • Data analysis skills, statistics are helpful
  • Good writing/communication skills

This project is an extension of a project to understand how people in different countries  think about anonymity on the internet. (CHI paper available if you are interested).


Project name: Urban Analytics

The convergence of social media, smartphones, and machine learning offers the opportunity to understand human behaviors at a fidelity and scale that previously was not possible. As one part of this vision, our team has been crawling geotag social media data (e.g. Twitter, foursquare, instagram) and applying machine learning techniques to understand various aspects of cities, e.g. how people perceive neighborhoods and what people do in these various neighborhoods. See our web site at livehoods.org for one example interactive visualization. We believe there are a large number of opportunities here, to help people with urban planning, business planning, transportation, resource allocation, and more.

We’re looking for undergrads to help us on this project, either by designing and evaluating different kinds of interactive tools that can help analysts understand this data, or by conducting new kinds of analyses to find interesting patterns. Useful skills for this project would include some subset of web programming, statistical analysis, information visualization, and/or machine learning, plus a lot of imagination.


Project name: CrowdScanner

Many of the apps on your smartphones are spyware. We’re building lots of tools that combine static analysis, dynamic analysis, and crowdsourcing to pinpoint unusual privacy- and security-related behaviors in apps and convey that to end-users in a meaningful way. One example of our analysis is the level of surprise over the behavior of apps: http://confabulator.blogspot.com/2012/11/analysis-of-top-10-most-unexpected.html

We’d like your help in (a) building better tools that combine crowdsourcing, static analyis, and dynamic analysis, and (b) creating web sites that the public (both end-users and developers) can use to understand and discuss unusual behaviors of their apps. Useful skills here include some subset of Java programming, machine learning, web design, web programming (databases, HTML, JavaScript, JSON), plus a lot of imagination.


Project name: Supporting Research & Investigation with Crowd-generated Timelines

We are exploring how crowdsourcing might provide valuable research assistance to scientists, journalists, historians, and others who frequently do research as part of their jobs. One project within this area involves building a tool that helps crowds create and visualize timelines, or lines of inquiry, for a given topic, person, or event. Researchers could then interact with, explore, and add to these timelines to help them better comprehend a research area and answer specific questions. Students will help design and develop this tool and assist in running experiments with workers in crowdsourcing platforms (e.g. Amazon Mechanical Turk) and professional researchers. The ideal candidate will have an interest in crowdsourcing and some experience with web development, visualization tools, and/or statistics and data analysis.


Project name: Starting New Online Communities

Online communities (e.g., Facebook, Live space, Wikipedia and open source projects such as Apache etc) have become an increasingly important element in our lives and also have opened up exciting opportunities for learning how these systems work. We are particularly interested in solving one question: what factors lead a new online community to continued success. We have two dataset: Wikia’s  four years archival data including 7601 Wiki-like communities; and DBLP’s compete data including thousands of conferences in computer science (each conference is an academic community). We plan to analyze the two datasets and answer the question of what factors lead to the success of new communities.

Relevant undergrad major or skills:

Major: Computer science or psychology

Skills: fluent in python/php/jave, fluent in SQL


Project name: Bosch Research & Technology Summer Internship  in Crowd-Powered System (Bosch R&T Centre & Carnegie Mellon University)

In a joint collaboration with Bosch Research and Technology Centre, CMU’s Social Computing Lab seeks enthusiastic and highly motivated interns to join the team and work on cutting edge crowd-powered systems. The internship will entail studying state-of-the-art crowd-sourcing systems in various domains and developing and testing a prototype system.

The intern will get an opportunity to work with Dr Niki Kittur at CMU and Bosch researchers; a successful internship will culminate in publications at a top HCI venue conference and patents.

Required knowledge and skills:

  • Currently pursuing a MS or PhD in CS, Social Computing and/or Data Mining
  • Strong programming experience in Java and comfortable with Web technologies
  • Ability to work independently
  • Excellent communication skills
Advertisements

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s