Online communities suffer serious newcomer attrition. This paper explores whether and how early activity diversity - the degree to which a newcomer engages in a wide range of a site's activities in the first session - is associated with their longevity. We introduce a metric (DSCORE) to characterize early activity diversity in online sites and run our analyses on an online community ‘MovieLens’. We find that DSCORE is significant both by itself and in conjunction with a measure of quantity of activity in predicting longevity. This finding is robust to different measures of longevity (aggregate number of sessions and attritions after sessions 1, 5, and 10). The immediate implication is an effective classifier for identifying users with higher (or lower) expected longevity from the first-session activity. We also find DSCORE is more useful than a traditional measure of measuring diversity such as the Gini-Simpson index. We conclude by discussing how early activity diversity may be more broadly effective in supporting design and management of online communities.

Prior work relevant to incorporating personality into recommender systems falls into two categories: social science studies and algorithmic ones. Social science studies of preference have found only small relationships between personality and category preferences, whereas, algorithmic approaches found a little improvement when incorporating personality into recommendations. As a result, despite good reasons to believe personality assessments should be useful in recommenders, we are left with no substantial demonstrated impact. In this work, we start with user data from a live recommender system, but study category-by-category variations in preference (both rating levels and distribution) across different personality types. By doing this, we hope to isolate specific areas where personality is most likely to provide value in recommender systems, while also modeling an analytic process that can be used in other domains. After controlling for the family-wise error rate, we find that High Agreeableness users rate at least 0.5 stars higher on a 5-star scale compared to low Agreeableness users. We also find differences in consumption in four different personality types between people who manifested high and low levels of that personality.

Striking up a good conversation with new acquaintances is often a difficult problem. In this paper we report on the perceptions of wearable device users who were given real-time personalized topic suggestions during a conversation with a person they just met. Suggestions were generated using a ranking recommendation algorithm, and were delivered via Google Glasses. We conducted a study with 38 pairs of strangers, who received such suggestions while conversing for the first time. Participants found the suggestions to be helpful, but only at the right moments, and for certain types of speakers. Our results contribute to the understanding of how communication interventions influence people's experience and behaviors, and enhance interpersonal interactions. Our study also presents design implications for applications on wearable devices to facilitate conversations between strangers.

Session identification is a common strategy used to develop metrics for web analytics and perform behavioral analyses of user-facing systems. Past work has argued that session identification strategies based on an inactivity threshold is inherently arbitrary or has advocated that thresholds be set at about 30 minutes. In this work, we demonstrate a strong regularity in the temporal rhythms of user initiated events across several different domains of online activity (incl. video gaming, search, page views and volunteer contributions). We describe a methodology for identifying clusters of user activity and argue that the regularity with which these activity clusters appear implies a good rule-of-thumb inactivity threshold of about 1 hour. We conclude with implications that these temporal rhythms may have for system design based on our observations and theories of goal-directed human activity.

In the era of big data, it is usually agreed that the more data we have, the better results we can get. However, for some domains that heavily depend on user inputs (such as recommender systems), the performance evaluation metrics are sensitive to the amount of noise introduced by users. Such noise can be from users who only wanted to explore the systems, and thus did not spend efforts to provide accurate inputs. Noise can also be introduced by the methods of collecting user ratings. In my dissertation, I study how user data can affect prediction accuracies and performances of recommendation algorithms. To that end, I investigate how the data collection methods and the life cycles of users affect the prediction accuracies and the performance of recommendation algorithms.

Eli Pariser coined the term 'filter bubble' to describe the potential for online personalization to effectively isolate people from a diversity of viewpoints or content. Online recommender systems - built on algorithms that attempt to predict which items users will most enjoy consuming - are one family of technologies that potentially suffers from this effect. Because recommender systems have become so prevalent, it is important to investigate their impact on users in these terms. This paper examines the longitudinal impacts of a collaborative filtering-based recommender system on users. To the best of our knowledge, it is the first paper to measure the filter bubble effect in terms of content diversity at the individual level. We contribute a novel metric to measure content diversity based on information encoded in user-generated tags, and we present a new set of methods to examine the temporal effect of recommender systems on the user experience. We do find that recommender systems expose users to a slightly narrowing set of items over time. However, we also see evidence that users who actually consume the items recommended to them experience lessened narrowing effects and rate items more positively.

One of the challenges for recommender systems is that users struggle to accurately map their internal preferences to external measures of quality such as ratings. We study two methods for supporting the mapping process: (i) reminding the user of characteristics of items by providing personalized tags and (ii) relating rating decisions to prior rating decisions using exemplars. In our study, we introduce interfaces that provide these methods of support. We also present a set of methodologies to evaluate the efficacy of the new interfaces via a user experiment. Our results suggest that presenting exemplars during the rating process helps users rate more consistently, and increases the quality of the data.

Most recommender systems assume user ratings accurately represent user preferences. However, prior research shows that user ratings are imperfect and noisy. Moreover, this noise limits the measurable predictive power of any recommender system. We propose an information theoretic framework for quantifying the preference information contained in ratings and predictions. We computationally explore the properties of our model and apply our framework to estimate the efficiency of different rating scales for real world datasets. We then estimate how the amount of information predictions give to users is related to the scale ratings are collected on. Our findings suggest a tradeoff in rating scale granularity: while previous research indicates that coarse scales (such as thumbs up / thumbs down) take less time, we find that ratings with these scales provide less predictive value to users. We introduce a new measure, preference bits per second, to quantitatively reconcile this tradeoff.