Data visualisation is something you see everywhere. On the news (tv, but also in news papers), social media, even on the street, in shops or on certain products. You might not realize it, but it’s everywhere around you. People try to visualize something to give you a better understanding of what’s happening. In data analytics (and probably everything data related) visualisation is important. I’ve mentioned it a couple of times before. You can do the most amazing analysis, find amazing results, but if you are not able to share this information properly, it loses all of it’s value.
In my opinion. data visualisation is something that is underestimated often. It really is hard to show a graph or table in a way that is immediately understandable for everyone. Why? Because you need to be able to visualize data in a way that everybody understands. As a data specialist, with a high interest in numbers, I look at numbers and graphs a lot. Even I often find it a challenge to figure out a plot or graph that I’m looking. But I’m trying to figure it out because I like data and I want to understand what (the hell) it’s saying. The majority of the people don’t have brains that work like that. People need to understand a plot within seconds, or they lose their interest. So, if you need to convince your manager, make sure he/she understands what he/she is looking at, otherwise it’s a lost game.
As a data analyst you need to step back from what you’ve been doing and try to scale down your (complex) findings in easy, small, acceptable chunks of information that is understandable. Make sure it levels with the person you present it to. They need to understand. This is so much harder and so much more important than that you think it is. This IS a deal breaker. You either make it, or break it in presenting this person your findings.
A graph should tell it’s own story. By just looking at it, you should understand what’s going on in that graph. It should not be necessary to explain the whole graph and it’s numbers. If that’s the case, it’s often not a good graph and something is not clear or too complex. A brain can only process so much. And if a brain can’t figure it out in about 5 seconds, it stops trying to understand, that’s the last thing you want. You’ll literally lose the person in your story trying to persuade them to do something. Easy to recognize if you lose them because their eyes often look as if they’ve been blinded by headlights (I bet you’ve seen that before). And the chances they’ll listen to you and actually support you in your findings are getting smaller and smaller.
Let’s take a look at dashboarding. Providing a dashboard for managers/customers so they have continuous insights in their data is important. The set up of this dashboard often is underestimated in time and complexity by clients and data analysts. Have you ever asked a client what it wants to see in its dashboard? It’s likely they kinda stare at you for a second or two, trying to come up with something and then mention something like revenue, purchases, or something like that. From a client perspective I totally understand this. It has probably never crossed your mind. Think about it: how would you know what you want to see in your dashboard, just from the top of your head?
From a data analytics perspective, it kinda leaves you in the dark. Yes, you can come up with all sorts of data related focus points and you can build a 16 pages long dashboard with every little data related detail in it.
- Is that effective? No.
- Is anyone gonna go through all that? No.
- Should you as a data analyst be able to pick out the most important things? Yes.
- But do you know exactly what’s important for a manager? No.
- Does he want top 10 sale numbers in a list, or a time series plot with sales over time? I don’t know.
That’s what it comes down to. It’s something only a manager is often able to answer. I don’t have THE solution for it, but in my experience it helps if you have a so called ‘dummy-dashboard’ in data studio for example. It doesn’t take too much time, but it helps visualize the data for the person that needs it. My experience so far is that every single person after seeing that is like:
- Oh, actually, can you show this in a graph over time?
- Oh, can you add this number?
- I’m only interested in the total, can you just get that for me.
You see? Because you gave it a little bit, their brain starts working and they suddenly get an idea of what they want. And that’s what you need, because now you can start building something for them that’s actually useful for them. And this process is the reason why it often takes more time to build something that’s useful instead of just building something.
Plots in dashboards
In my opinion a plot only needs to have information that is useful. Which means:
- Minimum amount of information, just enough (it’s a tricky one),
- Minimum distraction (no need for all colors of the rainbow in one graph),
- Focus on what matters,
- Get rid of the background grid if it doesn’t add anything,
- Name your graph, name your x and y ax properly,
- Highlight the line, dot, plot, bar, that is most important,
- Use correct numbers on the x & y ax (percentage, 1 or 2 decimal, same axes if you show two graphs to compare).
- Don’t add two graphs in one.
- Last but not least 1: don’t use circle diagrams EVER
- Last but not least 2: don’t use 3D
I know, it sounds like a lot, but all of this adds to the understandability of a graph. In my opinion it is truly worthwhile thinking about.
I’m always open for questions, discussions, feedback and other opinions.