Take one relationship: a plant 4 cm tall grows 1.5 cm per week. In words, that is the sentence you just read. As an equation, h=4+1.5w. As a table: week 0, 4 cm; week 2, 7 cm; week 4, 10 cm. As a graph: a straight line starting at height 4 and rising 1.5 for each week. All four describe the same plant. None of them is 'the' relationship; each is a way of writing it down.
Each format makes different questions easy. 'How tall after 10 weeks?' is quickest from the equation: 4+1.5×10=19 cm. 'When does it pass 12 cm?' can be read off the graph: just after week 5, since 4+1.5×5=11.5 and 4+1.5×6=13. 'Is the growth steady?' shows up in the table as equal jumps of 3 cm every 2 weeks. Switching formats is a problem-solving move, not decoration.
A representation re-encodes an idea that already exists; it doesn't change what the idea is. That separates it from abstraction, which builds a new idea by stripping many cases down to what they share. Redrawing h=4+1.5w as a graph is representation. Noticing that the plant, a taxi fare and a phone bill all follow 'starting value plus constant rate' is abstraction.
Think of it like this. Representations are like an architect's drawings of one house: a floor plan, a side view and a 3D sketch. The floor plan shows room sizes, the side view shows heights, and the sketch shows how the house will look, yet all describe the same building. A wrong drawing can mislead you, but the house itself doesn't change.
Core idea. A representation encodes a mathematical idea in a format — diagram, table, equation, or graph.