Data Science Applications in Public Health

Study Course Description

Course Description Statuss:Approved

Course Description Version:1.00

Study Course Accepted:15.02.2023 11:40:27

Study Course Information
Course Code:		SVUEK_145		LQF level:		Level 7
Credit Points:		2.00		ECTS:		3.00
Branch of Science:		Clinical Medicine; Health Care		Target Audience:		Health Management; Public Health
Study Course Supervisor
Course Supervisor:		Uģis Kārlis Sprūdžs
Study Course Implementer
Structural Unit:		Department of Public Health and Epidemiology
The Head of Structural Unit:
Contacts:		Riga, 9 Kronvalda boulevard, svekrsu[pnkts]lv, +371 67338307
Study Course Planning
Full-Time - Semester No.1
Lectures (count)		15	Lecture Length (academic hours)		1	Total Contact Hours of Lectures		15
Classes (count)		9	Class Length (academic hours)		1	Total Contact Hours of Classes		9
Total Contact Hours								24
Study course description
Preliminary Knowledge:
Research methods, basic statistics, preferably a mathematical understanding of logarithms and differentiation, computer literacy
Objective:
Acquire in-depth knowledge, understanding, and skills in specific methods of mathematical statistics and recently developed data science techniques for academic use and work in Public Health; also to promote the learning of data science terminology and its practical applications
Topic Layout (Full-Time)
No.	Topic			Type of Implementation		Number	Venue
1	Time series description and examples. Univariate time series analysis: trend, stationarity, seasonality, outliers.			Lectures		2.00	computer room
				Classes		1.00	computer room
2	Time series forecasting models: univariate forecasting. Trend and seasonality adjustments. Differencing. Outliers and one-time effects. Autoregression. Model evaluation. Notation. Forecasting interval.			Lectures		2.00	computer room
				Classes		1.00	computer room
3	Multivariate time series models. Structural vs. forecasting models. Concurrent effects. Time-lagged effects. Missing variables. Spurious correlation. Logarithmic models.			Lectures		2.00	computer room
				Classes		1.00	computer room
4	Multivariate time series models as a simulation platform.			Lectures		1.00	computer room
				Classes		1.00	computer room
5	Time series models compared with traditional SIR epidemiological models			Lectures		1.00	computer room
				Classes		1.00	computer room
6	Predictive models vs. descriptive models, definitions. Model architecture. Decision tree form as example of classification model. Model requirements.			Lectures		2.00	computer room
				Classes		1.00	computer room
7	Classification model development: data preparation and partitioning, model algorithm selection, validation. XGBoost method.			Lectures		2.00	computer room
				Classes		1.00	computer room
8	Classification model implementation. Identification of primary predictors and describing their contribution to the model. Model monitoring.			Lectures		2.00	computer room
				Classes		1.00	computer room
9	Brief overview of alternative modeling approaches. AutoML procedures.			Lectures		1.00	computer room
				Classes		1.00	computer room
Assessment
Unaided Work:
Independent work outside the classroom involves preparing for lectures, utilizing lecture notes to study for quizzes, doing homework assignments – data preparation, model development, and model assessment. Completing the study course evaluation questionnaire.
Assessment Criteria:
Active participation in class discussions and activities; Quizzes on the practical use of terminology and methods learned; Evaluation of homework assignments. Final exam covering terminology, methods, and practical applications – 40% Homework assignments involving data preparation and model development – 30% Quizzes – 30% Missed lectures require a minimum one-page summary of the covered material based on the information provided during the lecture.
Final Examination (Full-Time):		Exam
Final Examination (Part-Time):
Learning Outcomes
Knowledge:		Upon successfully completing the course students will: Understand time series analysis terminology and its implementation; Be familiar with time series analysis functionality in the OxMetrics package; Learn learn how to formulate, develop, and implement predictive classification models using the KNIME platform
Skills:		Upon successfully completing the course students will * Know how to open, create, and manipulate time series data in OxMetrics * Using OxMetrics, know how to prepare a proper descriptive univariate time series model * Using OxMetrics, know how to prepare a proper descriptive multivariate time series model * Using KNIME, know how to open data sets and prepare data for the development of predictive classification models * Know how to set up and execute a predictive model in KNIME * Know how to evaluate a predictive model in KNIME * Using KNIME, know how to identify the major drivers of a predictive model and their effect on the target variable * Know how to explain the implementation and monitoring of a classification model * Know how to summarize the methods taught in the course and their outcomes
Competencies:		Upon successfully completing the course students will Correctly interpret and evaluate applications of time series models in the Public Health field; Using Public Health data, plan, develop, and evaluate a predictive classification model
Bibliography
No.	Reference
Required Reading
1	Sprūdžs U. Lekciju materiāli 2022/2023
2	Russell S & Norvig P. Artificial Intelligence: A Modern Approach. Pearson, 2021.
3	Ārvalstu studentiem/For international students:
4	Russell S & Norvig P. Artificial Intelligence: A Modern Approach. Pearson, 2021.