Measuring Total Data Quality

U-M Free Access Enrollment Options

What You'll Learn

Learn metrics for evaluating Total Data Quality.
Create a quality concept map of TDQ from a particular application or data source.
Identify relevant software and related tools for computing the various metrics.

4 Modules

8 Hours

2 hrs per module (approx.)

Rating

About Measuring Total Data Quality

By the end of this second course in the Total Data Quality Specialization, learners will be able to:
1. Learn various metrics for evaluating Total Data Quality (TDQ) at each stage of the TDQ framework.
2. Create a quality concept map that tracks relevant aspects of TDQ from a particular application or data source.
3. Think through relative trade-offs between quality aspects, relative costs and practical constraints imposed by a particular project or study.
4. Identify relevant software and related tools for computing the various metrics.
5. Understand metrics that can be computed for both designed and found/organic data.
6. Apply the metrics to real data and interpret their resulting values from a TDQ perspective.

This specialization as a whole aims to explore the Total Data Quality framework in depth and provide learners with more information about the detailed evaluation of total data quality that needs to happen prior to data analysis. The goal is for learners to incorporate evaluations of data quality into their process as a critical component for all projects. We sincerely hope to disseminate knowledge about total data quality to all learners, such as data scientists and quantitative analysts, who have not had sufficient training in the initial steps of the data science process that focus on data collection and evaluation of data quality. We feel that extensive knowledge of data science techniques and statistical analysis procedures will not help a quantitative research study if the data collected/gathered are not of sufficiently high quality.

This specialization will focus on the essential first steps in any type of scientific investigation using data: either generating or gathering data, understanding where the data come from, evaluating the quality of the data, and taking steps to maximize the quality of the data prior to performing any kind of statistical analysis or applying data science techniques to answer research questions. Given this focus, there will be little material on the analysis of data, which is covered in myriad existing Coursera specializations. The primary focus of this specialization will be on understanding and maximizing data quality prior to analysis.

Skills You'll Gain

What You'll Earn

Certificate of Completion:: Certificates of completion acknowledge knowledge acquired upon completion of a non-credit course or program.

Experience Type: 100% Online
Format: Self-Paced
Series: Total Data Quality
Subject: Data Science

Technology
Platform: Coursera

Welcome Message

Welcome to Measuring Total Data Quality, a focused course designed to help learners assess and enhance the reliability and accuracy of data. This course covers key dimensions of data quality, including validity, origin, processing, access, source, missingness, and analysis. By completing this course, you will gain practical skills to evaluate data quality across different contexts, ensuring robust and trustworthy data for decision-making.

This abbreviated syllabus description was created with the help of AI tools and reviewed by staff. The full syllabus is available to those who enroll in the course.

Course Schedule

Module 1: Introduction and Measuring Validity and Data Origin Quality

Video: Welcome to Course 2!
Reading: Course Syllabus
Reading: Course Pre-Survey
Video: Measuring Validity for Designed Data
Reading: Files for Example 1
Video: Example 1: Performing CFA and Examining Measurement Invariance in R
Reading: Example 2: A tutorial on estimating 'true-score' multitrait-multimethod models with lavaan in R
Video: Approaches and Considerations for Measuring Quality for Gathered Data
Video: Measuring Validity for Gathered Data
Video: Measuring Data Origin Quality for Designed Data
Reading: Output and R data file for the next Examples video
Video: Examples: Computing Measures of Data Origin Quality for Designed Data in R
Reading: Case Study: Measuring the Quality of Cause-of-Death Data at the CDC
Video: Measuring Data Origin Quality for Gathered Data
Video: Example 4: Measuring Validity and Data Origin Quality for Gathered Data
Graded: Interpreting Validity Metrics
Graded: Interpreting Data Origin Quality Metrics

Module 2: Measuring Processing and Data Access Quality

Video: Measuring Processing Quality for Designed Data
Reading: Data files for the next example
Video: Example: Computing Processing Metrics with Real Data and Code
Video: Measuring Processing Quality for Gathered Data
Video: Example: Computing Processing Metrics for Gathered Data
Video: Measuring Data Access Quality for Designed Data
Video: Example: Computing Access Metrics with Read Data and Code
Video: Measuring Data Access Quality for Gathered Data
Reading: Case study article: Hino and Fahey 2019
Video: Case Study: Measuring Data Access Quality in Gathered Twitter Data
Graded: Interpreting Processing Metrics
Graded: Interpreting Access Metrics

Module 3: Measuring Data Source Quality and Data Missingness

Video: Measuring Data Source Quality for Designed Data
Reading: Data files for the next example
Video: Example: Computing Data Source Metrics with Real Data and Code
Video: Measuring Data Source Quality for Gathered Data
Video: Example: Computing Data Source Quality Metrics with Real Data and Code
Video: Measuring Threats to Data Source Quality: Designed Data
Reading: Link to R software and Examples on GitHub (from previous lecture)
Video: Example: Computing Data Missingness Metrics with Real Data and Code
Video: Measuring Data Missingness for Gathered Data
Reading: Data file for the next example
Video: Example: Computing Data Missingness for Gathered Data
Graded: Interpreting Data Source Quality Metrics
Graded: Interpreting Data Missingness Metrics

Module 4: Measuring the Quality of Data Analysis

Video: Measuring the Quality of an Analysis of Designed Data
Reading: Files for the next Example
Video: Example: Computing Measures of Data Analysis Quality for Designed Data in R
Video: Measuring the Quality of an Analysis of Gathered Data
Reading: Suggested readings from the previous lecture
Video: Example: Computing Metrics for Quality of Models of Gathered Data
Reading: The Aequitas Bias Toolkit for Auditing Machine Learning Models
Reading: Course Conclusion
Reading: References for Measuring Total Data Quality
Graded: Examining Analysis Quality Metrics and Interpreting Output

Grading Policy

This course includes a total of seven graded quizzes, each is worth between 14% and 16% of your final grade. Each quiz must be passed with a score of 80% or higher. Quizzes can be attempted once every four hours.