Distinguished CIDSE Seminar: Explorations of a Data Scientist in Three Multi-Disciplinary Domains

Louiqa Raschid, Professor at Smith School of Business & the Institute for Advanced Computer Studies, University of Maryland

Friday, April 26th 2013, 2:00 p.m. – 3:00 p.m.


“Explorations of a Data Scientist in Three Multi-Disciplinary Domains”

Abstract: A decade of activity in Linked Open Data has resulted in a wealth of datasets in the biological and life sciences. Our research has applied a range of algorithmic methods including dense subgraphs, graph summarization and bipartite match to explore annotated graphs – scientific concepts tagged with controlled vocabulary terms from ontologies.  We exploit ontological relatedness and shared annotations to identify interesting patterns that can summarize existing knowledge and potentially lead to discoveries.  We validate our results using a variety of ground truth including the sentence imprint in the literature and the phylogenetic relationships among genes.

Social data provides a rich source of individuals actions and interactions.  Based on research in diffusion and influence as well as features of user behavior reflecting the creation and curation of content, we develop models to predict the impact of user actions in engaging attention and diffusing a message. Our research also explores a hybrid network of followers, re-tweets and mentions on Twitter. We identify a novel heuristic that can outperform more expensive matrix factorization methods to make personalized (future) recommendations.

My presentation will conclude with a summary of my activities as an advocate for applying computational methods to better manage financial eco-systems. I will present a shared vision of data science for finance and a plan to develop community financial cyber infrastructure.

Bio:  Louiqa Raschid is a professor at the University of Maryland. She received her education at Bishop’s College and St. Bridget’s Convent in Sri Lanka, IIT Madras (B. Tech 1980) and the University of Florida (Ph.D. 1987).  Louiqa Raschid, together with her collaborators, has been addressing the challenges of data science for the past two decades. Her research explores the complexity of handling Linked Open Data, Web data, BIGDATA and streams of social data.

Raschid has led the Sahana FOSS project for disaster information management, serving as chief database architect and founding director of the Sahana Software Foundation. Sahana was initiated in the aftermath of the 2003 tsunami. It is the only comprehensive product for disaster information management that supports sharing of disaster data using open ontologies, standards and protocols and most recently it has supported Occupy Sandy.