Field Of Data Science Research Articles

Data Visualization Saliency Model: A Tool for Evaluating Abstract Data Visualizations.

Evaluating the effectiveness of data visualizations is a challenging undertaking and often relies on one-off studies that test a visualization in the context of one specific task. Researchers across the fields of data science, visualization, and human-computer interaction are calling for foundational tools and principles that could be applied to assessing the effectiveness of data visualizations in a more rapid and generalizable manner. One possibility for such a tool is a model of visual saliency for data visualizations. Visual saliency models are typically based on the properties of the human visual cortex and predict which areas of a scene have visual features (e.g. color, luminance, edges) that are likely to draw a viewer's attention. While these models can accurately predict where viewers will look in a natural scene, they typically do not perform well for abstract data visualizations. In this paper, we discuss the reasons for the poor performance of existing saliency models when applied to data visualizations. We introduce the Data Visualization Saliency (DVS) model, a saliency model tailored to address some of these weaknesses, and we test the performance of the DVS model and existing saliency models by comparing the saliency maps produced by the models to eye tracking data obtained from human viewers. Finally, we describe how modified saliency models could be used as general tools for assessing the effectiveness of visualizations, including the strengths and weaknesses of this approach.

The Best States for Data Innovation

The Best States for Data Innovation

System dynamics simulations for data-intensive applications

Simulation modeling is increasingly perceived as a methodological asset to the field of data science. Nonetheless, adequate graphical database interfaces are missing especially for most System Dynamics (SD) simulation tools. SimSyn is freely available middleware used to link together models developed in VENSIM with a PostgreSQL database for spanning SD models over geographic or multi-dimensional parameter space. The capabilities of SimSyn are demonstrated by simulating terrestrial carbon storage for 10,000 years on a 5 arc-min raster mesh with 278,115 grid cells. Results indicated the reasonable performance of data-linked simulations (7, 500 to 10,000 runs per 15 min) and a considerable increase of computational overheads associated with additional time series inputs. Apart from increasing performance, the incorporation of SD interoperability standards is a key objective for the further development of SimSyn.

An Introductory Guide to Data Science: The Terminological Landscape

The emerging field of data science has rapidly evolved into an extremely diverse field equipped with multi-disciplinary techniques to extract, analyze and classify structured and unstructured data. These methods offer researchers, policy analysts, and the lay public evidence-based insights into a tremendous range of human, organizational, and societal activities on a scale and scope that has rarely been possible with conventional scientific methods. At present, however, the multi-disciplinary nature of the data science space suffers a ‘language’ problem insofar as data scientists from different fields often use different terms to describe common methods and concepts. The aim of the present research is threefold. First, we report results of a literature review that identifies and defines the essential content domain of data science, with special focus on the classification of data collection techniques. Second, we establish a preliminary set of relationships among the most trafficked terms of data science to facilitate interdisciplinary communication among scientists from heterogeneous fields. And third, we develop a classification scheme of web-scraping methods based on their availability, the quality of the data procured by the method, the ease of data extraction, reproducibility, the technical skills required to leverage each method, and the types of data collected by each method.

Developing a checklist to inform data linkage study designs for health technology assessments: a case study linking the Cardiac Rhythm Management (CRM) register to the Secure Anonymised Information Linkage (SAIL) Databank

ABSTRACTObjectivesHealth services researchers are increasingly engaging with the emerging field of data science, but relatively few have the expertise to understand how innovative data linkage methodologies can, and cannot, be successfully applied in practice. There is little published guidance written specifically for this purpose.
 We aimed to develop study design criteria to help researchers in considering whether these methods are appropriate for their projects. A secondary objective was to test the criteria in a case study, and evaluate the application of the data linkage approach.
 ApproachClinical procedures requiring further research (according to the National Institute for Health and Care Excellence) were assessed against newly-developed CINDER criteria (Coverage; Identifiers; Numbers; Data; Existing records; Retrieval) to check the suitability of using data linkage methods. The CALON (Cardiac Ablation: Linking Outcomes for NICE) study was then established to evaluate outcomes of cardiac ablation procedures.
 Records from the UK’s Cardiac Rhythm Management (CRM) register were linked to routinely-recorded primary care, secondary care and mortality data in the Secure Anonymised Information Linkage (SAIL) Databank. Demographic profiles of patients identified from the register were compared with a group identified from SAIL. Outpatient attendances before and after ablation were compared using a Generalised Linear Mixed Model, assuming significance of p<0.05. Kaplan-Meier analyses were used to estimate survival.
 Evaluation of methodological success concentrated on: adequately characterising patient populations from existing data; matching individuals between datasets; and data completeness/consistency.
 ResultsThe linked dataset contained 2220 anonymised records. Almost all (99.7%) patients from the register were matched using deterministic and probabilistic techniques. These patients were similar to individuals identified from hospital records with respect to sex and comorbidity score (p>0.05); mean age differed by 1.6 years (95% CI 0.23-3.04; p=0.02).
 Patients accessed 26.7% fewer hospital outpatient appointments after ablation (95% CI 23.4 to 29.8; p<0.001). There was no significant difference in the number of primary care events before and after ablation (95% CI -4.3 to 4.0; p=0.91). Survival was estimated at 91.0% after 5 years. Insufficient granularity of data precluded subgroup analyses.
 ConclusionsData linkage can be used to evaluate outcomes of interventions, although there are limitations associated with secondary use of observational data. The CALON study identified a post-ablation reduction in hospital attendances, suggesting an overall improvement in general health.
 We aim to publish the full CINDER checklist as a generic resource to facilitate assessment of the feasibility of using data linkage methods in other projects.

Clinical Text De-identification and Other Large Scale Processing Tasks in Resource Constrained Environments

ABSTRACT
 ObjectivesClinical text de-identification is a common requirement of the ‘enclave’ governance model of ethical EHR research. However, there is often little consideration of the engineering task that is required to scale these approaches across the hundreds of millions of clinical documents containing personal identifiers that are resident in the data repositories of a typical NHS Trust. Similarly, natural language processing is an increasingly important field of clinical data science, yet it requires fault tolerant approaches to data processing. This work concerns the development of “turbo-laser” - a distributed document processing architecture based upon the popular ‘battle hardened’ Spring Batch framework - an industry standard for large scale processing tasks.
 ApproachUsing Spring Batch, we developed a highly scalable unstructured data processing framework, using the concept of remote partitioning. Remote partitioning allows us to offload processing tasks to any and all computers in a network. With this approach, it is possible to harness the entire compute available of an organisation, whether it be an office of 15 desktop PCs that go unused overnight, or a compute cluster of a thousand processors. This method is especially valuable in the NHS, where the provision of sufficient compute to make large scale analytics possible are often hindered by the lack of available hardware, or difficulties in navigating technical governance policies ill equipped for the demands of modern data science.
 ResultsTurbo-laser was developed in consideration of the processing challenges common in the NHS. Currently, four types of ‘job’ are available - De-identification, using the Cognition algorithm, generic GATE output, text extraction from binary files such as MS Office, PDF and scanned documents, and a document re-compiler to deal with EHR legacy issues. Examples of turbo-laser usage include processing 9 million binary documents on modest hardware, within 48 hours.
 ConclusionTurbo-laser is an enterprise grade processing tool, in keeping with the software engineering pattern of ‘batch processing’ that has been at the forefront of the informatics movement. An open source project, it is hoped that others may contribute and extend its principles, lowering the barrier of large scale data processing throughout the NHS.

Data Science in Libraries

EDITOR'S SUMMARYThe new field of data science involves advanced knowledge in statistics and computer science, combined with copious amounts of data. A report from the Big Data and Research Initiative under the Obama Administration, The Federal Big Data Research and Development Strategic Plan, calls attention to the roles that librarians will play in the future of data science. However, there are skills and management gaps librarians face that inhibit their ability to move forward in data science. A number of educational programs are now offered to remedy this problem, such as the Data and Visualization Institute for Librarians from North Carolina State University, the volunteer‐led Library Carpentry program, and most recently, the Data Sciences in Libraries Project, funded by the IMLS. This project aims to get librarians and library managers together to discuss the world of data science and create a roadmap for strategic planning.

Multilevel Data Processing Using Parallel Algorithms for Analyzing Big Data in High-Performance Computing

The growing gap between users and the Big Data analytics requires innovative tools that address the challenges faced by big data volume, variety, and velocity. Therefore, it becomes computationally inefficient to analyze such massive volume of data. Moreover, advancements in the field of Big Data application and data science poses additional challenges, where High-Performance Computing solution has become a key issue and has attracted attention in recent years. However, these systems are either memoryless or computational inefficient. Therefore, keeping in view the aforementioned needs, there is a requirement for a system that can efficiently analyze a stream of Big Data within their requirements. Hence, this paper presents a system architecture that enhances the working of traditional MapReduce by incorporating parallel processing algorithm. Moreover, complete four-tier architecture is also proposed that efficiently aggregate the data, eliminate unnecessary data, and analyze the data by the proposed parallel processing algorithm. The proposed system architecture both read and writes operations that enhance the efficiency of the Input/Output operation. To check the efficiency of the proposed algorithms exploited in the proposed system architecture, we have implemented our proposed system using Hadoop and MapReduce. MapReduce is supported by a parallel algorithm that efficiently processes a huge volume of data sets. The system is implemented using MapReduce tool at the top of the Hadoop parallel nodes to generate and process graphs with near real-time. Moreover, the system is evaluated in terms of efficiency by considering the system throughput and processing time. The results show that the proposed system is more scalable and efficient.

Thinking with Data: How to Turn Information into Insights, by Max Shron

As the emerging field of data science permeates many domains inclusive of information security and privacy, understanding the process of turning information into insights continues to be a major en...

Pervasive Computing Revisited

In 2001, Mahadev Satyanarayanan (Satya) published the paper, Pervasive Computing: Vision and Challenges, in IEEE Personal Communications. This was the first paper to substantively explore the implications of Mark Weiser's 1991 vision for a real-world systems infrastructure. It provided the conceptual foundations for what much of the computing industry now calls the Internet of Things (IoT). It also developed a taxonomy showing the deep relationships between distributed systems, mobile computing, and pervasive computing (IoT). This special issue is a retrospective on Satya's vision. The theme articles explore cyber foraging, energy efficiency, privacy concerns, and the emerging field of pervasive data science.

Pervasive Data Science

Technology is increasingly enabling us to instrument our physical environment with complex sensors and actuators, creating a connected world that generates huge volumes of complex data. In this article, the authors describe a number of opportunities and challenges that this new data-rich world brings to pervasive computing and highlight the emergence of a new field of pervasive data science. This article is part of a special issue on pervasive computing revisited.

A new data science research program: evaluation, metrology, standards, and community outreach

This article examines foundational issues in data science including current challenges, basic research questions, and expected advances, as the basis for a new data science research program (DSRP) and associated data science evaluation (DSE) series, introduced by the National Institute of Standards and Technology (NIST) in the fall of 2015. The DSRP is designed to facilitate and accelerate research progress in the field of data science and consists of four components: evaluation and metrology, standards, compute infrastructure, and community outreach. A key part of the evaluation and measurement component is the DSE. The DSE series aims to address logistical and evaluation design challenges while providing rigorous measurement methods and an emphasis on generalizability rather than domain- and application-specific approaches. Toward that end, each year the DSE will consist of multiple research tracks and will encourage the application of tasks that span these tracks. The evaluations are intended to facilitate research efforts and collaboration, leverage shared infrastructure, and effectively address crosscutting challenges faced by diverse data science communities. Multiple research tracks will be championed by members of the data science community with the goal of enabling rigorous comparison of approaches through common tasks, datasets, metrics, and shared research challenges. The tracks will permit us to measure several different data science technologies in a wide range of fields and will address computing infrastructure, standards for an interoperability framework, and domain-specific examples. This article also summarizes lessons learned from the data science evaluation series pre-pilot that was held in fall of 2015.

Data Scientist

The last installment of this column dealt with the specter of IT-caused unemployment. Here, the author considers a new IT-created employment opportunity--the data scientist. He looks at data, information, and knowledge and current IT job classifications to provide context, describes how big data has inspired the field of data science, and defines what data science is and what data scientists do.

A novel approach to information fusion in multi-source datasets: A granular computing viewpoint

The advent of Big Data has seen both the sources and volumes of data increase rapidly. A multi-source information system can be used to represent information drawn from multiple sources. However, some of these sources are of less importance than others, and some are essentially worthless. Selecting the most valuable sources and efficiently fusing information are therefore core issues in the field of data science. To investigate this matter, we first propose internal-confidence and external-confidence degrees to estimate the reliability of each information source within a multi-source information system. A source selection principle is then constructed, allowing worthy and reliable information sources to be chosen. Furthermore, a new information fusion method is constructed by transforming the original information of each object into a triangular fuzzy information granule, and some uncertainty measures of this fusion process are studied. Finally, to interpret and comprehend the proposed theories and algorithm, extensive experiments are performed on six datasets to verify that our approach can deal with practical issues. The results indicate that the proposed triangular fuzzy granule fusion approach is efficient and effective for information fusion in multi-source datasets.

Big Data Comes to School

The prospect of “big data” at once evokes optimistic views of an information-rich future and concerns about surveillance that adversely impacts our personal and private lives. This overview article explores the implications of big data in education, focusing by way of example on data generated by student writing. We have chosen writing because it presents particular complexities, highlighting the range of processes for collecting and interpreting evidence of learning in the era of computer-mediated instruction and assessment as well as the challenges. Writing is significant not only because it is central to the core subject area of literacy; it is also an ideal medium for the representation of deep disciplinary knowledge across a number of subject areas. After defining what big data entails in education, we map emerging sources of evidence of learning that separately and together have the potential to generate unprecedented amounts of data: machine assessments, structured data embedded in learning, and unstructured data collected incidental to learning activity. Our case is that these emerging sources of evidence of learning have significant implications for the traditional relationships between assessment and instruction. Moreover, for educational researchers, these data are in some senses quite different from traditional evidentiary sources, and this raises a number of methodological questions. The final part of the article discusses implications for practice in an emerging field of education data science, including publication of data, data standards, and research ethics.

Machine Learning with Big Data An Efficient Electricity Generation Forecasting System

Abstract Machine Learning (ML) is a powerful tool that can be used to make predictions on the future nature of data based on the past history. ML algorithms operate by building a model from input examples to make data-driven predictions or decisions for the future. The growing concept “Big Data” has brought much success in the field of data science; it provides data scalability in a variety of ways that empower data science. ML can also be used in conjunction with Big Data to build effective predictive systems or to solve complex data analytic problems. In this work, we propose an electricity generation forecasting system that could predict the amount of power required at a rate close to the electricity consumption for the United States. The proposed scheme uses Big Data analytics to process the data collected on power management in the past 20 years. Then, it applies a ML model to train the system for the prediction stage. The model can forecast future power generation based on the collected data, and our test results show that the proposed system can predict the required power generation close to 99% of the actual usage. Our results indicate that the ML with Big Data can be integrated in forecasting techniques to improve the efficiency and solve complex data analytic problems existing in the power management systems.

Informatics and data science: an overview for the information professional

Purpose – This paper aims to describe the emerging field of data science, its significance in the larger information landscape and some issues that distinguish the problems of data science and informatics from traditional approaches in the information sciences. Design/methodology/approach – Through a general overview of the topic, the author discusses some of the major aspects of how work in the data sciences and informatics differ from traditional library and information science. Findings – Data science and informatics, as emerging fields, are expanding our understanding of how the massive amount of information currently being generated can be collected, managed and used. While these may not be traditional “library” problems, the contributions of the library and information science communities are critical to help address aspects of these issues. Originality/value – The emerging fields of data science and informatics have not been extensively explored from the perspective of the information professional. This paper is designed to help information professionals better understand some of the implications of data science in a changing information environment.

Natural Language Search Interfaces: Health Data Needs Single-Field Variable Search.

BackgroundData discovery, particularly the discovery of key variables and their inter-relationships, is key to secondary data analysis, and in-turn, the evolving field of data science. Interface designers have presumed that their users are domain experts, and so they have provided complex interfaces to support these “experts.” Such interfaces hark back to a time when searches needed to be accurate first time as there was a high computational cost associated with each search. Our work is part of a governmental research initiative between the medical and social research funding bodies to improve the use of social data in medical research.ObjectiveThe cross-disciplinary nature of data science can make no assumptions regarding the domain expertise of a particular scientist, whose interests may intersect multiple domains. Here we consider the common requirement for scientists to seek archived data for secondary analysis. This has more in common with search needs of the “Google generation” than with their single-domain, single-tool forebears. Our study compares a Google-like interface with traditional ways of searching for noncomplex health data in a data archive.MethodsTwo user interfaces are evaluated for the same set of tasks in extracting data from surveys stored in the UK Data Archive (UKDA). One interface, Web search, is “Google-like,” enabling users to browse, search for, and view metadata about study variables, whereas the other, traditional search, has standard multioption user interface.ResultsUsing a comprehensive set of tasks with 20 volunteers, we found that the Web search interface met data discovery needs and expectations better than the traditional search. A task × interface repeated measures analysis showed a main effect indicating that answers found through the Web search interface were more likely to be correct (F 1,19=37.3, P<.001), with a main effect of task (F 3,57=6.3, P<.001). Further, participants completed the task significantly faster using the Web search interface (F 1,19=18.0, P<.001). There was also a main effect of task (F 2,38=4.1, P=.025, Greenhouse-Geisser correction applied). Overall, participants were asked to rate learnability, ease of use, and satisfaction. Paired mean comparisons showed that the Web search interface received significantly higher ratings than the traditional search interface for learnability (P=.002, 95% CI [0.6-2.4]), ease of use (P<.001, 95% CI [1.2-3.2]), and satisfaction (P<.001, 95% CI [1.8-3.5]). The results show superior cross-domain usability of Web search, which is consistent with its general familiarity and with enabling queries to be refined as the search proceeds, which treats serendipity as part of the refinement.ConclusionsThe results provide clear evidence that data science should adopt single-field natural language search interfaces for variable search supporting in particular: query reformulation; data browsing; faceted search; surrogates; relevance feedback; summarization, analytics, and visual presentation.

Segmentation of Factories on Electricity Consumption Behaviors Using Load Profile Data

In recent years, the new achievements in the field of technology and data science allowed to gather detailed and well-structured information about electricity consumption behaviors of industrial enterprises. Such type of information can find numerous applications in the power distribution industry. The utilities often use the data from contracts to assign each industrial customer a class label according to this type defined in predetermined industry segmentation. Such type of fixed-chart segmentation is not able to satisfy the needs of modern enterprises for the flexible and dynamic determination of production modes. In this paper, we address this problem by proposing a new method for the segmentation of various types of factories based on their electricity consumption patterns represented in load profile data. It exploits the evolution-based characteristics of smart meter data of multiple types of factories to remove irrelevant features. We use data visualization to estimate the number of clusters and apply the well-known $k$ -means algorithm on filtered data to generate segmentation. Experimental results on real load profile data collected with smart meters from manufacturing industries in Guangdong province of China have shown that the new clustering approach produced the meaningful segmentation of factories that reflect production operations.

Integrating Systems Modelling and Data Science

Although System Dynamics modelling is sometimes referred to as data-poor modelling, it often is –or could be– applied in a data-rich manner. However, more can be done in the era of ‘big data'. Big data refers here to situations with much more available data than was until recently manageable. The field of data science makes big(ger) data manageable. This paper provides a perspective on the future of System Dynamics with a prominent place for bigger data and data science. It discusses different approaches for dealing with bigger data. It reviews methods, techniques and tools for dealing with bigger data in System Dynamics, and sheds light on the modelling phases for which data science is most useful. Finally, it provides several examples of current applications in which big data, data science, and System Dynamics modelling and simulation are being merged.