Have you ever really thought about what kind of chart to use when you’re graphing data?
True Confession: I never did.
I just went on instinct and blind luck until I began working and learning as a data analyst. Now, I definitely do realize how valuable visualizations are for communicating data. But I worry that the value is often missed, and even totally mishandled sometimes because we just don’t know any different.
And I know why – realistically most of us who are doing data analysis in the credit union industry are learning as we go. Very few of us have degrees that focus on in Business Intelligence, Database Engineering, or formal Data Visualization (DataViz) skills.
But every single one of us can know that there is a difference between a pie chart and a bar graph even when each might use the exact same data.
Or, when a stacked column is actually useful (or is it just another version of a pie?)
This topic came up recently as I spoke with a new analyst at a credit union where the board has expressed a preference for visualizations rather than tables of data. We had a good conversation about how some types of charts are better for showing information over time, others are better for showing small differences in volume from one thing to another, etc.
But truthfully, although I was speaking as if I knew (because now I do), I had only recently gained confidence and knowledge of the definitive strategies on my own.
I’m no expert yet by far – but I have been doing a bit of reading lately to teach myself some of these most common chart and graph strategies. As a data analyst I do consider it my responsibility to make sure I’m a master of my craft and I need to KNOW (not just go on instinct) that I am organizing and communicating information in ways that are helpful to my audience.
This is one of those things that people don’t often realize makes a difference until you do it right. But once you have a clean, logical visual….people just “get it” and smoothly move into thinking about the results rather than still trying to interpret what the results are even saying in the first place.
The most important points on my mind today are these realizations that:
- Not all charts are created equal
- A bad (confusing) chart will hurt your analyst brand and mislead decision-makers.
The decision to make about which chart is the right one for your data depends on many things. Don’t let the endless variations and strategies overwhelm – Begin reading a little bit about the general ideas where each type is valuable. I think you’ll find the dots begin connecting (haha) very quickly.
Good luck! To start I highly recommend The Data Viz Project – http://www.datavizproject.com/