Understanding how to separate the structure and semantics of data when creating models is essential to moving to a datacentric environment. The syllabus for each model is listed below.
Data, Information, and Knowledge:¶
Transforming data into knowledge requires understanding the importance of and how to collect all of the additional meaning of the data. Sometimes this is called metadata. In other instances, it is called context or semantics. In this lesson, we establish processes to collect this information and discuss the importance of making this information computable.
Types of data vs. datatypes:¶
There exists some confusion about the importance of the distinction between the two terms datatypes and types of data. In this lesson, we clarify these terms and then discuss how to analyze any real-world situation where a domain expert might want to create a computable model for some concept.
In this lesson, we discuss how data standards are created and why there are complications with the implementation of many of them.
Ontologies & Controlled Vocabularies:¶
Ontologies and controlled vocabularies are almost never complete nor are they universally accurate. However, they do provide a useful tool in our ability to share meaning. In this lesson, we discuss how and when to use these tools as well as discover some of the caveats to their use.
Determining Concept Granularity¶
When beginning the process of datacentric modeling, the first step is to determine what concepts you are going to model. In this lesson, we use the Kunteksto tool to create some basic data models with semantics.
Modeling Existing Forms¶
Many times the concepts we wish to model are already defined by paper or web forms. In this lesson, we explore creating data models from existing forms. This lesson goes in depth in utilizing what was learned about datatypes and converting what we see on the form into the correct datatypes, the constraints and user interface options.
Datacentric Tools (1)¶
In this lesson, we take the above work and translate it into executable models using the Datacentric Tools Suite.