Take "fraction" as a node and draw links of three kinds. Dependence: fractions need division and equal parts. Analogy or equivalence: 43 is the same value as 0.75 and 75%, and it behaves like a probability of 3 favourable outcomes out of 4. Application: slope (runrise), unit prices ($6 for 4 kg is $1.50 per kg) and scale drawings. With that many links, fractions sit at the centre of a busy part of the map.
Links matter because they let methods travel. The chance of drawing 2 red marbles from a bag of 3 red and 2 blue can be found with a tree diagram, with a product of fractions 53⋅42=103, or by counting pairs, 3 red pairs out of 10. Three routes through the network reach the same answer, so if one is blocked another still works. Some links are surprising: 1+3+5+7=16 is also a picture of a 4-by-4 square built from L-shaped layers, tying number patterns to geometry.
A concept network is not a prerequisite chain. "You need fractions before ratios" is one arrow of one kind. The chain tells you what order to learn things in; the network also records sideways links (analogies) and outward links (applications), which is what lets you use an idea somewhere new. If a map has only required-before arrows, it is a dependency chart, a small part of the network.
Think of it like this. A concept network is a subway map: stations are ideas and lines are links. Some stations sit on many lines, so you can reach almost anywhere from them, and when one line is closed you transfer to another. Unlike subway lines, links come in different kinds (needs, resembles, used in), and each kind takes you somewhere different.
Core idea. A concept network is the web linking math ideas by dependence, analogy, and application.