Yesterday I saw this graph trending on Social Media. Many people were celebrating the fact that the US Government’s actions seemed to be tanking bitcoin. 
Indeed, that precipitous drop at the end there is impressive. But what struck me immediately is that the y axis starts at 80,000, not at 0. For comparison, I made the same graph in google sheets:
** Note that the values in my google sheets version are slightly different, because I used web plot digitizer to extract the data from the graph (I couldn’t find the original data easily), and the process is not precise.
and then I edited it so that the Y axis starts at 0.
It displays exactly the same numbers, but tells a rather different story!
You’ll note there are no labels on the X axis, because I could not read them on the original. When you look at the text on the original screenshot, you can see that this is the stock price over a mere 24 hours. From 5pm on April 5th to 5pm on April 6th.
Where you start (and, indeed, finish) the Y axis makes a huge difference to the story the graph tells.
Incidentally, you can achieve a similar effect by extending the upper limit of the y axis, like here, where I’ve made the top value 500000.
To be clear: All four graphs are technically correct. There is no right or wrong place to start the Y axis. In this case, since the change took place over a single day, and it’s actually an unusually large change for a stock price over such a short period of time, starting from 80,000 highlights the dramatic nature of the change, so it makes sense. It all depends on what story you’re trying to tell with the data.
Inspired by my friend Sarah, here is a different way to show the same information.
This time it’s not exactly the same numbers. Instead, I’ve calculated the percentage change from the initial price, using a very simple formula.
(price-first)/first*100)
This is a really good way to show the change in value. It wanders around a lot, going up a bit at the start, but that last drop is really clear. And the values are easier for us to wrap our heads around than big numbers like 80,000. What is the difference between 80000 and 83000? It’s hard to grasp. A 3% drop, though, is much more intuitively meaningful. I can get a grip on that.
** Note that the percentage change in the original screenshot is calculated from the closing price the day before, not the start of the graph
You can also change the story this graph tells by changing the scale, of course. Here I’ve dropped the minimum value from -4 to -100.
So if I wanted to make people think that the stock price had not changed much, which graph would I use?
Once again, it’s important to stress that all of these graphs are technically accurate. It’s important, when you’re making graphs, to think about the story you want to tell with the data. What type of graph, and what features of the graph, will help you tell that story?
Likewise, when you’re looking at someone else’s graph, we all need to apply that critical data literacy and look at the scale on the y axis, as well as checking the labels, finding out the origin of the data, and considering whether the graph is an accurate, valid way to display that data, and what story it might be trying to tell. You might just avoid being taken in by fake news!
Remember that you can support ADSEI’s work in data literacy by donating at givenow.com.au/adsei/
