Sukjin You

Measuring happiness of US cities by mining user-generated text in Flickr.com: a pilot analysis

This poster describes a methodology to numerically represent the happiness of a city by mining user generated terms in Flickr.com. As a pilot analysis, we collected 15,000 text records consisting of titles, tags, descriptions, and comments for the thirty most populous cities in the United States. Parsed text was utilized to calculate happiness scores (H-Score) by matching text extracted from Flickr.com with a happiness index dictionary. In addition, we examined the relationships between the calculated H-scores and real world phenomena including population, crime rate, and climate. Based on this pilot analysis, a future study is planed that involves a large dataset with prediction analysis.

Is informal data citation for data sharing and re-use more common than formal data citation?: Is Informal Data Citation for Data Sharing and Re-Use More Common than Formal Data Citation?

Data citation to reflect instances of data sharing and re‐use is becoming more common, although it is not yet widely adopted. We investigate how common formal and informal data citation are in bioscience/biomedical research. We found that informal data citation (i.e., acknowledgments embedded in publications outside of formal references) is far more common than formal citations. Informal citations go unrecorded, and therefore authors of datasets largely do not receive credit for having informed or influenced the research of others.

Situational Virtual Reference: Get Help When You Need It

This study aims to increase the use of virtual reference service by increasing the awareness of the availability of the service to users who really need it. A new situationally-based virtual reference interface, called the sVR interface, has been designed to reflect different levels of user search success. Findings from an eight-month field study done in a university library improved our understanding of how to effectively enhance the availability of virtual reference service to users who need it. A discussion about balancing the availability and the intrusiveness of virtual reference service is also provided.

Informal data citation for data sharing and reuse is more common than formal data citation in biomedical fields

Data citation, where products of research such as data sets, software, and tissue cultures are shared and acknowledged, is becoming more common in the era of Open Science. Currently, the practice of formal data citation—where data references are included alongside bibliographic references in the reference section of a publication—is uncommon. We examine the prevalence of data citation, documenting data sharing and reuse, in a sample of full text articles from the biological/biomedical sciences, the fields with the most public data sets available documented by the Data Citation Index (DCI). We develop a method that combines automated text extraction with human assessment for revealing candidate occurrences of data sharing and reuse by using terms that are most likely to indicate their occurrence. The analysis reveals that informal data citation in the main text of articles is far more common than formal data citations in the references of articles. As a result, data sharers do not receive documented credit for their data contributions in a similar way as authors do for their research articles because informal data citations are not recorded in sources such as the DCI. Ongoing challenges for the study of data citation are also outlined.

Using Event Identification Algorithm (EIA) to Improve Microblog Retrieval Effectiveness

Different from traditional Web page or database document, microblog has its unique features. Considering its sensitivity to the time, we introduce a new factor to improve the tweet retrieval effectiveness. The ranking score of a retrieved tweet is adjusted by the event identification algorithm (EIA). We evaluated our approach using 49 search topics and a total of 9,096,198 tweets collected provided by the trec 2011 microblog track. Our initial results indicated that EIA helped to improve the performance for topics with worse original average precision (AP) score and for topics that are represented with relatively short queries (one or two terms).

Visualized related topics (VRT) system for health information retrieval

To help bridge the gap between consumer users vocabulary and controlled vocabulary used to index health information, in this demo we implemented a Visualized Related Topics (VRT) browser system. The VRT was integrated into the “MeshMed” [2] system to support health information retrieval. The key technology behind the VRT browser is to select MeSH terms, which represent the related topics or subjects, from the top relevant documents. We rank these MeSH terms using the traditional Term Frequency-Inverse Document Frequency (TF-IDF) algorithm. The VRT browser displays a graphic representation of these MeSH terms by creating a visual where the selected MeSH terms stem from the centered user query. The design goal is provide users an overview of the key topics of the search results. In addition, VRT browser may also help users form better queries. Using the VRT browser we will be studying how to effectively assist in consumer users with their health information seeking.

Balancing factors affecting virtual reference services: identified from academic librarians' perspective

Many digital libraries are providing Virtual Reference Services (VRS). There could be various approaches to increase the quality of VRS. In this study, we focused on two key factors; improving helpfulness and reducing users feeling of intrusiveness. Studies indicated that librarian-initiated attempts for help may increase users feeling of intrusiveness [2] [3]. It is challenging to provide high helpfulness along with less intrusiveness in VRS. This study aimed to identify factors that contribute to improving helpfulness and reducing intrusiveness. Data were collected based on a survey using systemic random sample approach. Our initial results indicated that awareness, timing, and transparency were key factors affecting the helpfulness and intrusiveness.