Learning Graph for FanPedia: An Intelligent Textbook on Fans & Blowers¶
Open Learning Graph Viewer Fullscreen
This section contains the learning graph for this textbook. A learning graph is a graph of concepts used in this textbook. Each concept is represented by a node in a network graph. Concepts are connected by directed edges that indicate what concepts each node depends on before that concept is understood by the student.
A learning graph is the foundational data structure for intelligent textbooks that can recommend learning paths. A learning graph is like a roadmap of concepts to help students arrive at their learning goals.
At the left of the learning graph are prerequisite or foundational concepts. They have no outbound edges. They only have inbound edges for other concepts that depend on understanding these foundational prerequisite concepts. At the far right we have the most advanced concepts in the course. To master these concepts you must understand all the concepts that they point to.
This book's graph has 581 concepts and 1,098 dependency edges in 13 categories. Its concepts were drawn from Twin City Fan's FanPedia: The Fan & Blower Encyclopedia and the current product categories on tcf.com.
Here are other files used by the learning graph.
Course Description¶
We use the Course Description as the source document for the concepts that are included in this course. The course description uses the 2001 Bloom taxonomy to order learning objectives.
List of Concepts¶
We use generative AI to convert the course description and its sources into a Concept List. Each concept is a short Title Case label of at most 32 characters.
Concept Dependency List¶
We next use generative AI to create a Directed Acyclic Graph (DAG). DAGs do not have cycles where
concepts depend on themselves. We provide the DAG in two formats. One is a CSV file and the other
format is a JSON file that uses the vis-network JavaScript library format. The vis-network format uses nodes, edges and metadata
elements with edges containing from and to properties. This makes it easy for you to view and edit the learning
graph using an editor built with the vis-network tools.
Analysis & Documentation¶
Course Description Quality Assessment¶
This report rates the overall quality of the course description for the purpose of generating a learning graph.
- Course description fields and content depth analysis
- Checks that the course description has enough depth for 300-600 concepts
- Identifies content gaps and strengths
- Suggests areas of improvement
View the Course Description Quality Assessment
Learning Graph Quality Validation¶
This report gives you an overall assessment of the quality of the learning graph. It uses graph algorithms to look for specific quality patterns in the graph.
- Graph structure validation: all 581 concepts are in one connected graph
- DAG validation: no cycles detected
- Foundational concepts: 3 entry points (Air Movement, Fan Speed, Horsepower)
- Indegree distribution analysis
- Longest dependency chain: 18 concepts
View the Learning Graph Quality Validation
Concept Taxonomy¶
In order to see patterns in the learning graph, it is useful to assign colors to each concept based on the concept type. We use generative AI to create about a dozen categories for our concepts and then place each concept into a single primary classifier.
- 13 categories, from air and fan fundamentals through catalogs, selection and specification
- Category organization: foundational elements first, specification and the capstone last
- Every category holds between 5% and 10% of the concepts, well under the 30% threshold
- Clear 3-5 letter abbreviations for use in the CSV file
Taxonomy Distribution¶
This report shows how many concepts fit into each category of the taxonomy. Our goal is a somewhat balanced taxonomy where each category holds a similar number of concepts. We also don't want any category to contain over 30% of our concepts.
- Statistical breakdown
- Detailed concept listing by category
- Visual distribution table
- Balance verification