Part-time Faculty, Supply Chain and Business Technology Management
Conferences and presentations
Dr. Dzieciolowski’s career combines 30 years of experience in data science and leadership roles in the Canadian telecommunication industry with teaching and research, as a part-time faculty member at John Molson School of Business in Montreal.
In his professional career, he led Data Science and Modelling teams in Rogers, Fido and Bell Canada where he developed predictive models and analytical solutions for key business problems.
Currently, Dr Dzieciolowski serves as a Chief Analytics Officer at Daesys Inc., the startup company developing Machine Learning Operations (MLOps) platform for Machine Learning models.
At JMSB, Dr. Dzieciolowski researches and teaches courses in data mining, statistics, data science and analytical software to undergraduate and graduate students. He has built synergies between industry and academia and provided internship funding for students of Montreal universities. He has developed and launched a joint Certification in Business Analytics and Predictive Modelling between JMSB and SAS – the world’s leader in data science software.
Dr. Dzieciolowski research interests focus on statistics and data science methods; in particular, on models diagnostics and explainability, predictive models, multivariate analysis, forecasting, social networks and text analytics. He was distinguished with the CUPFA Special JMSB Research award for the project in big data in 2016.
He obtained M.Sc. in mathematics from Warsaw University in Poland and Ph.D. in statistics from Queen’s University in Kingston, Ontario.
He is an active promoter of data science and a frequent speaker at academic and industry conferences.
The objective of the course is to gain a basic understanding of the concepts, principles and techniques of data management and statistical analyses through the use of Statistical Software. The focus of this course is on practical aspects of data management and statistical analysis with the help of Statistical Analysis System (SAS) – widely used by business, government and academia. In the process of familiarizing with SAS, principal attention will be given to data management techniques related to handing of data storage, information retrieval, variable and row transformation, data cleansing, file handling and preparation of data for analytics, visualization, statistical analysis and reporting. The course helps students understand methods of handling big data through the use of principles of efficient programming. Through hands-on approaches, students learn how to design and execute programs for the data-intensive processes.
The objective of the course is to introduce basic concepts and techniques of Data Mining and its applications in business. In recent years, Data Mining, often also referred to as Data Science, has become one of the most dynamically growing areas of business knowledge discovery field. Proliferation of computing resources, Internet, social networks, mobile communication with the resulting deluge of data, often called -Big Data, has motivated companies to turn data into an asset for decision-making and used gain competitive advantage.
The course focuses on supervised and unsupervised data mining methods, their underlying principles and applications for structured data as well as unstructured data such as text. The data mining techniques introduced throughout the course include regression models and learning algorithms such as k-nearest neighbours, naïve Bayes, classification and regression trees, support vector machine, neural networks, discriminant analysis, association rules, cluster analysis and text mining.
The course focuses on the theory andapplications of the most used methods of business forecasting, includingsmoothing, differencing, exponential smoothing, decomposition, linearregression models with varying data patterns, as well as ARMA, ARIMA and seasonalARIMA models. Some more advancedmethods are explored at the end of the course.Key features of time seriesdata are explored. The course also introduces tools needed to evaluate many forecasting scenarios:identifying benchmark forecasting methods, diagnostic checking to determine ifa method has adequately utilized the available information in the data,techniques for computing prediction intervals, and methods for evaluating forecastaccuracy. Practical usage of the methods and techniquesare demonstrated through caseassignments and a term project, usingbusiness and economic datasets.Students are expected to have a basic knowledge of statistical concepts tofollow this course.
Thiscourse emphasizes the development of analytical skills needed to workeffectively in a business environment. It introduces the basics of knowledgediscovery from big data, business analytics and predictive modelling for datadriven decision-making and policy formation. Topics include exploratory dataanalysis, statistical analysis, and modelling, forecasting and datavisualization. Practical usages of the methodologies are demonstrated viaprojects and case analysis based on big data from various functional areas ofbusiness.
Towards Explainable Machine Learning Operations (MLOps)
Classifier Rank - A New Classification Assessment Method
Explaining Machine Learning Models
Explaining Black-box Machine Learning Models
K. Dzieciolowski, "Creating a Successful Data Science Program – A Joint Academic and Industry Perspective"
K. Dzieciolowski, "Fuzzy Matching and Predictive Models for Acquisition of NewCustomers"
Predicting Customer Attrition with Markov Chains
Predictive Probabilities in Marketing
Why Data Mining is Difficult to Learn? And What Can We Do to Make It Easier?
Data Mining in Marketing Acquisition Campaigns
K. Dzieciolowski, R. Ducharme, Y. Bengio, A. Lukban, NameMining for Long Distance Calling Pattern Recognition.
© Concordia University