Scott Counts

Identifying topical authorities in microblogs

Content in microblogging systems such as Twitter is produced by tens to hundreds of millions of users. This diversity is a notable strength, but also presents the challenge of finding the most interesting and authoritative authors for any given topic. To address this, we first propose a set of features for characterizing social media authors, including both nodal and topical metrics. We then show how probabilistic clustering over this feature space, followed by a within-cluster ranking procedure, can yield a final list of top authors for a given topic. We present results across several topics, along with results from a user study confirming that our method finds authors who are significantly more interesting and authoritative than those resulting from several baseline conditions. Additionally our algorithm is computationally feasible in near real-time scenarios making it an attractive alternative for capturing the rapidly changing dynamics of microblogs.

Predicting postpartum changes in emotion and behavior via social media

We consider social media as a promising tool for public health, focusing on the use of Twitter posts to build predictive models about the forthcoming influence of childbirth on the behavior and mood of new mothers. Using Twitter posts, we quantify postpartum changes in 376 mothers along dimensions of social engagement, emotion, social network, and linguistic style. We then construct statistical models from a training set of observations of these measures before and after the reported childbirth, to forecast significant postpartum changes in mothers. The predictive models can classify mothers who will change significantly following childbirth with an accuracy of 71%, using observations about their prenatal behavior, and as accurately as 80-83% when additionally leveraging the initial 2-3 weeks of postnatal data. The study is motivated by the opportunity to use social media to identify mothers at risk of postpartum depression, an underreported health concern among large populations, and to inform the design of low-cost, privacy-sensitive early-warning systems and intervention programs aimed at promoting wellness postpartum.

Supporting social presence through lightweight photo sharing on and off the desktop

Lightweight photo sharing, particularly via mobile devices, is fast becoming a common communication medium used for maintaining a presence in the lives of friends and family. How should such systems be designed to maximize this social presence while maintaining simplicity? An experimental photo sharing system was developed and tested that, compared to current systems, offers highly simplified, group-centric sharing, automatic and persistent people-centric organization, and tightly integrated desktop and mobile sharing and viewing. In an experimental field study, the photo sharing behaviors of groups of family or friends were studied using their normal photo sharing methods and with the prototype sharing system. Results showed that users found photo sharing easier and more fun, shared more photos, and had an enhanced sense of social presence when sharing with the experimental system. Results are discussed in the context of design principles for the rapidly increasing number of lightweight photo sharing systems.

Social media as a measurement tool of depression in populations

Depression is a serious and widespread public health challenge. We examine the potential for leveraging social media postings as a new type of lens in understanding depression in populations. Information gleaned from social media bears potential to complement traditional survey techniques in its ability to provide finer grained measurements over time while radically expanding population sample sizes. We present work on using a crowdsourcing methodology to build a large corpus of postings on Twitter that have been shared by individuals diagnosed with clinical depression. Next, we develop a probabilistic model trained on this corpus to determine if posts could indicate depression. The model leverages signals of social activity, emotion, and language manifested on Twitter. Using the model, we introduce a social media depression index that may serve to characterize levels of depression in populations. Geographical, demographic and seasonal patterns of depression given by the measure confirm psychiatric findings and correlate highly with depression statistics reported by the Centers for Disease Control and Prevention (CDC).

mimir: a market-based real-time question and answer service

Community-based question and answer (Q&A) systems facilitate information exchange and enable the creation of reusable knowledge repositories. While these systems are growing in usage and are changing how people find and share information, current designs are inefficient, wasting the time and attention of their users. Furthermore, existing systems do not support signaling and screening of joking and non-serious questions. Coupling Q&A services with instant and text messaging for faster questions and answers may exacerbate these issues, causing Q&A services to incur high interruption costs on their users. In this paper we present the design and evaluation of a market-based real-time Q&A system. We compared its use to a similar Q&A system without a market. We found that while markets can reduce wasted resources by reducing the number of less important questions and low quality answers, it may also reduce the socially conducive questions and usages that are vital to sustaining a Q&A community.

Major life changes and behavioral markers in social media: case of childbirth

We explore the harnessing of social media as a window on changes around major life events in individuals and larger populations. We specifically examine patterns of activity, emotional, and linguistic correlates for childbirth and postnatal course. After identifying childbirth events on Twitter, we analyze daily posting patterns and language usage before and after birth by new mothers, and make inferences about the status and dynamics of changes in emotions expressed following childbirth. We find that childbirth is associated with some changes for most new mothers, but approximately 15% of new mothers show significant changes in their online activity and emotional expression postpartum. We observe that these mothers can be distinguished by linguistic changes captured by shifts in a relatively small number of words in their social media posts. We introduce a greedy differencing procedure to identify the type of language that characterizes significant changes in these mothers during postpartum. We conclude with a discussion about how such characterizations might be applied to recognizing and understanding health and well-being in women following childbirth.

Pathfinder: an online collaboration environment for citizen scientists

For over a century, citizen scientists have volunteered to collect huge quantities of data for professional scientists to analyze. We designed Pathfinder, an online environment that challenges this traditional division of labor by providing tools for citizen scientists to collaboratively discuss and analyze the data they collect. We evaluated Pathfinder in a sustainability and commuting context using a mixed methods approach in both naturalistic and experimental settings. Our results showed that citizen scientists preferred Pathfinder to a standard wiki and were able to go beyond data collection and engage in deeper discussion and analyses. We also found that citizen scientists require special types of technological support because they generate original research. This paper offers an early example of the mutually beneficial relationship between HCI and citizen science.

Understanding affect in the workplace via social media

We investigate the landscape of affective expression of employees at a large Fortune 500 software corporation via an internal microblogging tool. We present three analyses of emotional expression among employees, based on literature in organizational behavior: its relationship to (1) exogenous/lifestyle and endogenous workplace factors, (2) geography and (3) organizational structure. We find that employees tend to make significant accommodations in affect expression when interacting with others over the organizational hierarchy. We also find that positive affect is expressed through interpersonal communications that connect disparate geographic regions. Our findings have implications for enabling emotional reflection of employees and for management in that they can help uncover emotional patterns associated with episodes of high and low productivity, allowing organizations to improve employee engagement and promote positive attitudes.

Group-Based Mobile Messaging in Support of the Social Side of Leisure

Communication on mobile devices plays an important role in people’s use of technology for leisure, but to date this communication has largely been one-to-one. Mobile internet connectivity can support a variety of group-based messaging and media sharing scenarios. Switching to group-based messaging should enhance the social and leisure aspects of the communication, but in what ways and to what extent? An experimental system for text and photo messaging on mobile devices was tested in a research deployment to four groups of 6–8 participants who used both a group-based and one-to-one version of the system. Results highlight a significant increase in message sending, in mobile device “fun”, and in the social qualities of mobile communication when messaging group-wide, along with a few minor costs. Qualitative feedback provides further explanation of the social benefits.

Exploring wearable ambient displays for social awareness

Mobile phones represent not only a means of communication, but an increasingly omnipresent computing platform, enabling diverse modes of communication including ambient displays that are tied to bodies and social groups, rather than physical environments. As an example of such a display, we present Damage, a prototype device for mobile ambient awareness of a social group, and discuss design considerations for such devices.