Reading a crowded graph

Often a graph carries a great deal of information and it is hard to know where to start. A two-step process helps.

Step 1: Identify what is given. Notice how the graph is organised, what scale is used, and what patterns the data shows. Do this before drawing any conclusion.

Take a graph of worldwide rocket launches by organisation. For each organisation, the launches for 2021, 2022 and 2023 are shown as three adjacent bars. The scale is 1 unit length = 20 rockets. An 'Others' category clubs many organisations together to keep the graph short.

One more thing to notice about ORDER. In the onion price graph the months ran January to December, because the order carries meaning — it shows change over time. Here, changing the order of the organisations would not change the meaning at all.

Step 2: Infer from what is given. Now analyse and interpret each observation. The USA, China and Russia are the leading rocket-launching countries in this period. SpaceX launched about twice as many rockets in 2022 as in 2021, and about 35 more in 2023 than in 2022.

Arianespace's number decreased every year. United Launch Alliance launched more in 2022 than 2021, but fewer in 2023 than in either. Other organisations launched about 25 rockets in 2023.

Then test claims against the data, and be strict about what it can support. 'All organisations launched more rockets than the previous years' is false — Arianespace fell every year. 'The average number of rockets launched by CASC in these 3 years is around 40' can be checked by reading the three bars and averaging. A good graph also leaves you curious: why does the USA launch so many more? What did earlier and later years look like?