Student and Faculty Research
Data Science Institute – Research Projects
Data Science has an impact on nearly every discipline. From our environment, to our food and health habits, from weather forecasting to predicting financial markets, from smart phone use to advertising strategies, Data Science has been a fast-evolving and ever-widening field. At Saint Peter’s, interdisciplinary research in data science and data analytics involves cutting edge topics, novel technology and data from industry, government and academic partners.
Research is an integral part of the Data Science graduate program. M.S. students engage in an interdisciplinary research project under the guidance of a faculty member for their Research Capstone Project.
Faculty research interests include big data, machine learning, artificial intelligence, quantitative methods, biomedical engineering, data visualization, consumer analytics, operational research, data mining, parallel computing, optimization, predictive modeling, informatics and health outcome research.
Analytics topics include business, health, finance, social media, marketing, consumer, politics, road safety, crime, manufacturing, geographic imaging, clinical and biomedical data to name just a few.
Below are some examples of current student and faculty research projects.
Business and Insurance Fraud Detection
In collaboration with a leading analytics company, we are analyzing a large body of transactional records and news articles to identify names and locations involved in insurance frauds. We use machine learning and natural language processing to perform efficient text mining tasks that self-learn information about crimes and fraud and aid in their detection.
We are working with multiple partners in this rapidly developing field. In one project, we analyze 24/7 recordings of heart pulse and breath rhythm to detect abnormal patterns such as sleep apnea. We are also investigating the application of using personal monitoring data to support large-scale clinical trials for pharmaceutical companies.
Business and Marketing Analytics
How to use data to formulate marketing, consumer and business strategy is becoming increasingly important for organizations. In this area several projects are being undertaken in a wide range of existing and emerging areas. For example, we are analyzing usage patterns of cell phone subscribers to predict churn between cell phone vendors. We are also investigating large-scale claim data to perform market research on self-driving cars in both the US and the EU. In addition, students are analyzing the effects of economic indicators on consumer equity using data from multiple sources.
Geographic Data and Road Safety
In collaboration with our industry partners, we identify in real-time the traffic patterns and probability of road accidents to present a route of maximum safety for the vehicle operator. We use Geographic Information Systems (GIS) software such as QGIS and ArcGIS, and high performance computing for real-time prediction.
Big Data and Social Media
Big data associated with social media and blog posts offers unprecedented opportunities to understand the public mood in politics, consumer preference and user sentiment. In one project, students are analyzing keywords from Twitter pages to conduct sentiment analysis for opinion mining and Internet marketing campaigns. We are also investigating algorithms to mine patterns from large-scale social media data.
Data Science and Financial Markets
Real-time visualization and analysis of stock prices or market data is critical for optimizing a portfolio. In this project we use Tableau Software and Python scripting to analyze market prices at opening and closing of NYC Stock Exchange. This project involves several collaboration partners.
Higher Education Data & Analytics
In this project, we analyze the flow of college applicants across States. Students selected features such as distance from home, tuition rates and local job demand to identify the potential reasons behind a brain drain between States.
Real Time Monitoring and Crime Analytics
Big data and crime has been an emerging topic. Students and faculty have been analyzing crime rate and prediction using statistical learning methods and crime data from Bogota, Colombia as well as cities in U.S. Working with industry, we also develop predictive methods for crime prevention in other cities and regions.
Data Mining and Visualization
Mining and visualizing patent data is important for business intelligence. In this project, students work with industry partners to explore leading-edge techniques for automating patent analysis and visualization using US Patent and Trademark Office data.
Data Science and Distributed Computing
In this project, we explore the development of novel technology and application of cloud computing to help process large amounts of data. Different technological platforms are compared based on their functionalities, offerings and accessibility for small and medium businesses. Students work in close collaboration with a corporate partner.
Biomedical Imaging and Clinical Informatics
Collecting, processing and visualizing hospital imaging data in real time has important benefits for medicine. In this project, students are exposed to real time analytics and clinical informatics used by hospital partners and clinicians.
Students are encouraged to publish in peer-reviewed journals and present their research at academic conferences.
With the Age of Big Data upon us, we risk drowning in a flood of digital data. Big data spans five dimensions (volume, variety, velocity, volatility and veracity), generally steered towards one critical destination - value. Big data has now become a critical part of the business world and daily life. Containing big information and big knowledge, big data does indeed have big value. IJDS confronts the challenges of extracting a fountain of knowledge from "mountains" of big data.
The International Journal of Business Analytics (IJBAN) is an indispensable resource for practitioners and academics that work in Business Analytics and related fields. Business Analytics is commonly viewed from three major perspectives: descriptive, predictive, and prescriptive. Business Analytics provides the framework to exploit the synergies among traditionally-diverse topics, such as the fields of data mining, quantitative methods, OR/MS, DSS, and so forth, in a more practical, application-driven format. The journal bridges the gap among different disciplines such as data mining, business process optimization, applied business statistics, and business intelligence/information systems. The journal supports and provides tools to allow companies and organizations to make frequent, faster, smarter, data-driven, and real-time decisions.
The International Journal of Information Systems and Supply Chain Management (IJISSCM) examines current, state-of-the-art research in the areas of SCM and the interactions, linkages, applications, and support of SCM using information systems. This journal encompasses theoretical, analytical, and empirical research, comprehensive reviews of relevant research, and case studies of effective applications in this area. The use of new technologies, methods, and techniques are emphasized.
The impacts of society touch everyone. They may be considerable, even critical. Societal problems, especially those remaining unchecked and unmanaged, can weaken, threaten, or even destroy, society. Production produces pollution; medication generates side-effects; pesticide causes poisoning; innovation can lead to unemployment, etc. IJSSoc deals not only with whether modern society should be sustainable, but also the ways in which this could/should come about, balancing economic development/environmental protection, real aggregate demand/aggregate supply, human beings/nature, consumption/preservation, material/spiritual pleasures, civil liberty/self-restraint, hedonism/practicality, science/society.
Management science should help different managers and policy makers to make optimal, if possible, or satisfactory, at least, informed decisions. IJAMS builds six interfaces: between management science theory and application; between management scientists and managers; between "hard" decision models and "soft" decision models; between deterministic and probabilistic models; between a specific strategy and an individual model; and finally, between a corporation and its environment, as well as the whole of society.
Facilitating transformation from data to information to knowledge is paramount for organizations. Companies are flooded with data and conflicting information, but with limited real usable knowledge. However, rarely should a process be looked at from limited angles or in parts. Isolated islands of data mining, modeling and management (DMMM) should be connected. IJDMMM highlights integration of DMMM, statistics/machine learning/databases, each element of data chain management, types of information, algorithms in software; from data pre-processing to post-processing; between theory and applications.
Many current data analysis techniques are beyond the reach of most managers and practitioners. Obscure maths and daunting algorithms have created an impassable chasm for problem solvers and decision makers. IJDATS bridges three gaps: firstly, a gap between academic ivory tower and the real world; secondly, a gap between quantitative data analysis techniques and qualitative data analysis techniques; and finally, a gap between a specific technique and an overall strategy.
The International Journal of Operations Research and Information Systems (IJORIS) examines current, state-of-the art advances in the interactions, linkages, applications, and support of operations research with information systems. Covering emerging theories, principles, models, processes, and applications within the field, this journal provides practitioners, educators, and researchers with an international collection of all operations research facets.
In today's fast-paced business environment, even with an abundance of information, decision-making can be complex and slow. As floods of data emerge, effective information processing is sought as a panacea. With the ever-present spectre of uncertainty, sound decisions are key. As a consequence of the various conflicts/dilemmas, employment of efficient data management leading to better decision-making is the goal. Organisations must employ effective information management/decision-making processes at each critical stage of their functions. IJIDS addresses the issues involved in this.
The International Journal of Information Systems in the Service Sector (IJISSS) examines current, state-of-the-art research in the area of service sectors and their interactions, linkages, applications, and support using information systems. This fully refereed journal encompasses theoretical, analytical, and empirical research, as well as comprehensive reviews of relevant research, technical reports, book reviews, and case studies of effective applications in this area. The use of new theories, technologies, models, methods, techniques, and principles are emphasized.