The most honest chart I ever made was the one I almost did not publish, because it refused to tell the story the editor wanted. We had three seasons of audience data and a headline already half-written about growth. The data, politely, declined to grow.
What it showed instead was stranger and better: the audience was not getting bigger, it was getting wider. Different people, different hours, a longer tail of the merely curious. The numbers had been there all along; the wrong chart had been hiding them behind a single rising line.
A visualization is an argument that happens to be made of pixels. If it argues a thing that is not true, the pixels are the least of your problems.
I think of data storytelling as photography with a longer exposure. You are still framing, still deciding what to leave out, still choosing the moment — it just unfolds over a season instead of a five-hundredth of a second. And the temptation is the same: to crop until the picture agrees with you.
The discipline is to resist that crop. Show the spread. Keep the season that breaks the trend. The reader can tell when a chart is performing confidence it has not earned, the same way they can tell when a portrait is performing emotion the subject did not feel.