The phrase: From Data to Information, and from Information to Knowledge, has become a cliché but it has never been as fitting as today. With the emergence of Big Data and the need to make sense of the massive amounts of disparate collection of individual datasets, there is a requirement for practitioners of data-driven domains to employ a rich set of analytic methods. Whether during data preparation and cleaning, or data exploration, the use of computational tools has become imperative. However, the complexity of underlying theories represent a challenge for users who wish to apply these methods to exploit the potentially rich contents of available data in their domain. In some domains, text-based data may hold the secret of running a successful business. For others, the analysis of social networks and the classification of sentiments may reveal new strategies for the dissemination of information or the formulation of policy.
My own research and that of my students falls in the domain of computational epidemiology. Designing and implementing tools that facilitate the study of the progression of diseases in a large population is the main focus in this domain. Complex simulation models are expected to predict, or at least suggest, the most likely trajectory of an epidemic. The development of such models depends on the availability or data from which population and disease specific parameters can be extracted. Whether census data, which holds information about the makeup of the population, of medical texts, which describe the progression of disease in individuals, the data exploration represents a challenging task. As many areas that employ data analytics, computational epidemiology is intrinsically multi-disciplinary. While the analysis of some data sources may reveal the number of eggs deposited by a mosquito, other sources may indicate the rate at which mosquitoes are likely to interact with the human population to cause a Dengue and West-Nile Virus epidemic. To convert information to knowledge, computational scientists, biologists, biostatisticians, and public health practitioners must collaborate. It is the availability of sophisticated visualization tools that allows these diverse groups of scientists and practitioners to explore the data and share their insight.
I first met Hector Cuesta during the Fall Semester of 2011, when he joined my Computational Epidemiology Research Laboratory as a visiting scientist. I soon realized that Hector is not just an outstanding programmer, but also a practitioner who can readily apply computational paradigms to problems from different contexts. His expertise in a multitude of computational languages and tools, including Python, CUDA, Hadoop, SQL, and MPI allows him to construct solutions to complex problems from different domains. In this book, Hector Cuesta is demonstrating the application of a variety of data analysis tools on a diverse set of problem domains. Different types of datasets are used to motivate and explore the use of powerful computational methods that are readily applicable to other problem domains. This book serves both as a reference and as tutorial for practitioners to conduct data analysis and move From Data to Information, and from Information to Knowledge.
Armin R. Mikler
Professor of Computer Science and Engineering
Director of the Center for Computational Epidemiology and Response Analysis
University of North Texas