Read the frame before the trend
A chart is a selection of information. Its title tells you what the author wants you to notice, but the labels tell you what you are actually seeing. Start with the unit, the place, the population and the period. A count of events is different from a rate per person. A monthly comparison answers a different question from a comparison with the same month a year earlier.
Consider an invented example: a service records 100 complaints one year and 120 the next. That is a 20 percent increase in complaints. If the number of customers also doubled, however, the complaint rate per customer fell. Both calculations can be correct. Which one matters depends on whether the question concerns total workload or the experience of an individual customer.
Ask what the picture makes easy to miss
Check where the vertical axis begins. A narrow range may help reveal small differences, but it can also make those differences look dramatic. Starting at zero is not automatically the right choice for every chart; the important point is whether the scale is clear and appropriate to the question. Look for changes in the spacing of the scale as well.
Then inspect the beginning and end of the time window. A line measured from an unusual peak can tell a different visual story from one measured over a longer period. That does not prove manipulation. It does mean the chosen window deserves a reason. Missing observations, revised figures and changes in definitions can also interrupt a clean-looking comparison.
Turn the graphic back into a sentence
Try describing the result without a metaphor: what changed, by how much, over which period, according to which data? Add the main limitation. If the sentence becomes less exciting but more precise, you have probably improved your understanding.
Keep a distinction between description and explanation. Two lines moving together do not, by themselves, show that one caused the other. A good graphic can reveal a question worth investigating without answering it. The aim is not to distrust every chart. It is to give the numbers enough context that the design cannot do all the reasoning for you.
This is an original educational guide, not a breaking-news report. Prepared with AI assistance. No interviews, on-the-ground reporting or firsthand experience are claimed.