Clemens Drews

Summarizing sporting events using twitter

The status updates posted to social networks, such as Twitter and Facebook, contain a myriad of information about what people are doing and watching. During events, such as sports games, many updates are sent describing and expressing opinions about the event. In this paper, we describe an algorithm that generates a journalistic summary of an event using only status updates from Twitter as a source. Temporal cues, such as spikes in the volume of status updates, are used to identify the important moments within an event, and a sentence ranking method is used to extract relevant sentences from the corpus of status updates describing each important moment within an event. We evaluate our algorithm compared to human-generated summaries and the previous best summarization algorithm, and find that the results of our method are superior to the previous algorithm and approach the readability and grammaticality of the human-generated summaries.

Home Location Identification of Twitter Users

We present a new algorithm for inferring the home location of Twitter users at different granularities, including city, state, time zone, or geographic region, using the content of users’ tweets and their tweeting behavior. Unlike existing approaches, our algorithm uses an ensemble of statistical and heuristic classifiers to predict locations and makes use of a geographic gazetteer dictionary to identify place-name entities. We find that a hierarchical classification approach, where time zone, state, or geographic region is predicted first and city is predicted next, can improve prediction accuracy. We have also analyzed movement variations of Twitter users, built a classifier to predict whether a user was travelling in a certain period of time, and use that to further improve the location detection accuracy. Experimental evidence suggests that our algorithm works well in practice and outperforms the best existing algorithms for predicting the home location of Twitter users.

Social Aspects of Using Large Public Interactive Displays for Collaboration

Large displays have several natural affordances that can simplify small group collaborative work. They are large enough to hold multiple work areas, they are easy to see and can be manipulated directly via touch. When placed into group and public spaces, such displays create pervasively available working surfaces for lightweight, temporary walkup use. The BlueBoard is a large plasma display with touch sensing and a badge reader to identify individuals using the board. The onboard software acts as a thin client giving access to each participants web-based content (e.g., home pages, project pages). The client also has a set of tools and mechanisms that support rapid exchange of content between those present. The overall design of the BlueBoard is one that is easily learnable (under 5 minutes), very simple to use, and permits novel uses for collaboration. Our initial field study revealed a number of social issues about the use of a large pervasively available display surface, yet indicates that a shared public display space truly has distinct properties that lend themselves to sharing content. Extreme learnability & overall simplicity of design makes BlueBoard a tool for collaboration that supports intermittent, but effective use for side-by-side collaboration between colleagues.

Virtual jukebox: reviving a classic

Recent advances in compression technology, combined with lower cost of storage and bandwidth, have made digital distribution of rich content including music not only technically feasible but also popular with a broad audience. However, limited progress has been made in the way this content is enjoyed by eyed users. We focus on the problem of playing music in a shared space, e.g, office, home, car - such that all listeners who are present share a positive music experience. Our scheme enables collaborative selection of content and pooling of content files. Users can express their preferences by contributing songs to be played and through a simple voting scheme. The system builds profiles and automatically selects content for playback, maximizing the match with the groups taste. As users vote, the system learns more about their collective preferences and can adjust the playlist accordingly, thus providing an incentive mechanism.

Recommending targeted strangers from whom to solicit information on social media

We present an intelligent, crowd-powered information collection system that automatically identifies and asks targeted strangers on Twitter for desired information (e.g., current wait time at a nightclub). Our work includes three parts. First, we identify a set of features that characterize ones willingness and readiness to respond based on their exhibited social behavior, including the content of their tweets and social interaction patterns. Second, we use the identified features to build a statistical model that predicts ones likelihood to respond to information solicitations. Third, we develop a recommendation algorithm that selects a set of targeted strangers using the probabilities computed by our statistical model with the goal to maximize the over-all response rate. Our experiments, including several in the real world, demonstrate the effectiveness of our work.

Exploring patterns of social commonality among file directories at work

We studied files stored by members of a work organization for patterns of social commonality. Discovering identical or similar documents, applications, developer libraries, or other files may suggest shared interests or experience among users. Examining actual file data revealed a number of individual and aggregate practices around file storage. For example, pairs of users typically have many (over 13,000) files in common. A prototype called LiveWire exploits this commonality to make file backup and restore more efficient for a work organization. We removed commonly shared files and focused on specific filetypes that represent user activity to find more meaningful files in common. The Consolidarity project explores how patterns of file commonality could encourage social networking in an organizational context. Mechanisms for addressing the privacy concerns raised by this approach are discussed.

Recent shortcuts: using recent interactions to support shared activities

We present an empirical study of teams that revealed the amount of extraneous individual work needed to enable collaboration: finding references to other people, finding files to attach to email, managing incoming email attachments, managing the variety of files used in shared activities, and tracking what work is owed to others. Much of this work involves finding recently accessed objects that are needed again in the users current task focus. These observations led to the design of Recent Shortcuts, a tool to help support coordination by making recently used objects easily accessible. Recent Shortcuts enables quick access to people (including groups of people), received attachments, files, and file folders that the user interacted with recently for re-use in the users current context. Recent Shortcuts makes it easy to use these objects across applications with no additional user input and minimal changes to the users applications or work practice. Early user experiences with a working prototype led to an extension that integrates recently accessed objects across multiple devices.

Unobtrusive but invasive: using screen recording to collect field data on computer-mediated interaction

We explored the use of computer screen plus audio recording as a methodological approach for collecting empirical data on how teams use their computers to coordinate work. Screen recording allowed unobtrusive collecting of a rich record of actual computer work activity in its natural work setting. The embedded nature of screen recording on laptops made it easy to follow the users mobility among various work sites. However, the invasiveness of seeing all of the users interactions with and through the computer raised privacy concerns that made it difficult to find people to agree to participate in this type of detailed study. We discuss measures needed to develop trust with the researchers to enable access to this rich, empirical data of computer usage in the field.

Retail applications of signature verification

The dramatic rise in identity theft, the ever pressing need to provide convenience in checkout services to attract and retain loyal customers, and the growing use of multi-function signature captures devices in the retail sector provides favorable conditions for the deployment of dynamic signature verification (DSV) in retail settings. We report on the development of a DSV system to meet the needs of the retail sector. We currently have a database of approximately 10,000 signatures collected from 600 subjects and forgers. Previous work at IBM on DSV has been merged and extended to achieve robust performance on pen position data available from commercial point of sale hardware, achieving equal error rates on skilled forgeries and authentic signatures of 1.5% to 4%.

Test-driven development for the web: increasing efficiency of web development

With the rapid growth of World Wide Web, demands on website developers have increased dramatically. At the same time new web development challenges have emerged. These challenges include enabling web developers with a low level of experience, fast paced development cycles and a disconnect between different phases of web development. In this paper we present algorithms which address some of these challenges. Our algorithms lower the barrier of expertise and experience required to transform development requirements into web pages, bridge the divide between web testing and development as well as improve test case maintenance. We also present a survey we conducted among web developers to understand their problems, experiments to demonstrate the performance of our algorithms and a user study that shows the value of our approach.