How I’m training others in improving their visualization skills

Do you want to help someone improve their visualization skills?
Or do you want to train others on any other statistics skills?

Training is a very important topic for us statisticians. We train people all the time it can be about statistics or non-statistics.

Today, I’ll be sharing some of my insights, tips, and tricks on how to do that in an effective way.

  1. Set expectations
    • What do they think they will get in the training?
    • What are their goals?
  2. Start with a case story
    • Show the value of visualization
  3. Speak about the difference between explanatory and exploratory data visualization
    • Where do they mostly work with?
    • Create a connection with the audience
  4. Speak through an example
    • Ask them to sketch the data by using pen and paper
    • Create an emotional trigger
    • Repetition improves quality
    • Provide self-evaluation
    • Supportive Feedback
    • Eliminating clutter
  5. Review the learnings
    • Reflecting on what they have learned

Listen to this episode and share this with your friends and colleagues!


Alexander: You’re listening to the effective statistician podcast, the weekly podcast with Alexander Schacht,  Benjamin Piske  and Sam Gardner, designed to help you reach your potential to lead great science and serve patients without becoming overwhelmed by work.

Subscribe to our Newsletter!

Do you want to boost your career as a statistician in the health sector? Our podcast helps you to achieve this by teaching you relevant knowledge about all the different aspects of becoming a more effective statistician.

Today, I’m talking about how I’m training others, in improving their visualization skills. I’ll give you all the hints and tips and tricks, how to do that in an effective way. So, stay tuned and listen to some music. I’m a big fan of PSI. That’s why I’m also promoting PSI here and I’m really doing this, not just because PSI is sponsoring this, but because PSI is really an awesome community. 

I’m producing this podcast in association with Psi. Psi is a community dedicated to Leading and promoting the use of Statistics within the healthcare industry for the benefit of Patience. There’s a lot you can get from PSI, there is video on demand Content Library, there’s free registration to PSI webinars. There are reduced fees for all kinds of different events and much, much more. And the fee is really small for non-high income countries. It is just 20 pounds, by the way for students It’s free. So if you know, any students tell them, it’s for free and it’s also only 95 pounds Annually for high-income countries that’s about the same number in Euros or in Dollars. At the current rate, visit the PSI website at PSI to learn more about PSI activities and become a PSI member to today.

Training is a very important topic for us statisticians. We are training people all the time. Today, I want to give you insights on how I train other statisticians and non-statisticians on data visualization. And there’s a couple of things that you will take from here that you can apply actually to any training. The first step is that I set expectations. I think this is really important because if you don’t set expectations, then people can easily be dissatisfied at the end. And also you need to make sure whether it’s the right training for them. I once went through the steps and one of the participants said, that’s not for me. I can save this 90 minutes of training, that’s completely fine. That’s much better than spending 90 minutes and then getting a bad recommendation and bad feedback because you didn’t set expectations. How do I set expectations? Well, I ask the participants about what they think they will get from the training. Of course, over the learning page and the  whole registration process, of course, you have already set expectations in terms of what this training is about. Who is the target audience? What are the goals and things like this, but it’s always good to kind of Here’s more specifically about it, when I’m in a face-to-face situation. I even write these expectations down on a whiteboard on a flip chart, so that I can at the end of the meeting or training, come back to that and show that kind of expectation 1 take-off we discussed about that expectation 2 take the off, we also discussed that expectation 3 mostly, but we didn’t completely get to that. Second thing, start with a case study. It’s always great to start with stories that engage people that show the value of data visualization. It’s especially great, If the case study is something that resonates with them. That’s why I really love to have homogeneous groups in terms of people that I train on data visualization. So, for example, if you want to train Physicians, it’s great to have people that come from the same indication. Let’s say, you would explain something to dermatologists. Then it’s great to have data visualization about psoriasis or atopic dermatitis or something like this and not something, In terms of Cardiology, whereas I don’t even really understand what this is about. 

These stories help to connect emotionally and that best works. If that is something that people can understand. So if you explain something about data visualization to your finance department, then come up with something about Financial, something that may be about the stock market or something about the what is the benefit of purchasing different things? Or how do you visualize where the different gains and losses in the business come from these kinds of things? And that leads directly into the point of why is visualization important? Why will it help them? What is important for them? That is also something where you can connect with why learning is now important for them and that will also help them to see how they will use it in the next way, as another step, as a next Step. I’m speaking about the difference between explanatory and exploratory data visualization, and I asked them after I’ve talked about this. Where do they mostly work? Is it either explanatory or exploratory or both? And you see, there’s another kind of Engagement? I’m trying to connect and interact with the participants as much as possible. This is especially important for virtual training, because, in Virtual environments you get so easily distracted. And asking them to interact even if it’s just writing as if it was a chat. I’m exploratory or explanatory both. That keeps the engagement time. Next step is then to get them actually doing something. So I speak through an example. Example data sets are relevant to them. And ask them to sketch this data with pen and paper. And then I did a little bit of a trick, so I asked them to do this. And I say, if you have any questions. Just let me know. I have done this a couple of times. So I wait for a couple of seconds and then usually nobody has any questions or maybe a little bit about the data set and then I say, could you stop for a moment? You don’t have all the information you need. Could you imagine what is missing here? Sometimes really clever. People are in it and directly recognize it, but very often, it takes some time up to the moment where they see. Hmm. Say I’m missing 0 information about the context. Who’s the audience? In which environment will the data visualization be used, will it be used on a smartphone or is it used in a big Auditorium? Is it used in paper or in a study report? What’s the goal that we want to achieve here, what specific message do we want to convey here? And so that helps them to really get this point home. This kind of little trick is really effective for them to remember context matters because if this is uncomfortable with them, and if this feeling uncomfortable, helps them to remember things better. I learned about that from some psychologists and they said you really learn best to remember if there’s emotional connection to something. And that emotional connection can be something bad or something good. But, there needs to be an emotional trigger that will help to point more easily to remember it. And so with this little trick, I put some on the spot that helps them to remember it better.

 Now even in virtual situations. I always ask them to do the sketching with pen and paper. Why pen and paper? Well, I also tell that in the training. Pen and paper have a couple of benefits. So the first is you’re not limited by how well you know the tool, be it Excel or art or SAS or whatsoever. You can be free in terms of sketching things, and also you can build it more naturally. You need to make every decision yourself. You don’t work from a template. If you start with Excel, you always start with a template. If you start with art  you always start with a template. Yes, with a default setting, and If you need to develop it yourself, you’re making conscious decisions about all these little things. The second is, much more likely to discard something on pen and paper. iteration is really important, trying out lots of different things is really important and with pen and paper, you can sketch out lots of different visualizations in a very short period of time. Whereas if you do it as an art, unless you’re a wizard in art? It always will take you quite some time to do it all in Excel. And once you have already spent, let’s say two hours on data visualization. You have sanction cost bias, you’re attached to this graphic already and you don’t want to throw it away because you have already invested so much. And to overcome that. It’s really good to lower the effort. So with pen and paper, you can find new things over time, which is really nice. And in this repetition, you increase the quality. 

 There’s also some nice research about it, if you are asked two groups to produce something and you’re asked one group to do it perfectly and the other group to have lots of quantity. The surprising result is that it is very often the second group. The one that is asked for quantity, actually develops better quality because over time, they learn and learn. And that way, instead of trying to get one thing really great. Through the process. They get it done really well because they always make decisions again and again and re-evaluate them all the time. In that way you come to a better result. I also talked about practicing or exercising,  repetition and seeking feedback quite a lot. So when people have done these sketches I speak about feedback and why it’s important. And I bring an example that even absolute gurus very often get it wrong the first time, for example, you can refer to Alberto Cairo who is training me. He often discards things after he has presented it to the targeted audience or completely needs to revise it. Because even though you might be an expert in data visualization, you never can completely anticipate how it will come across to the audience in a given context. So when everybody has sketched the data visualization, I usually give them something like 10 minutes for it. I asked them to provide, self evaluation of it. So I asked them to critique it themselves first. That shows me directly what they have already understood, what they didn’t do, because maybe they didn’t have any colored pencil or something like this, what I would have done differently and then I can say, Yes. Here’s a really nice principle that you applied. So some Gestalt principle, for example, or here you really eliminated clutter or something like this. So I can highlight certain good things. And I can also say, potentially you can do this here differently. Yeah, for example, if you have participants who have done vertical bar charts and the labels for each of these bars is pretty long and I can say, hey, why don’t you use the horizontal bar charts, and you have much more space to write your labels. So you can sell some suggestions? But also emphasize what they have done well and that will create an environment where there’s really good kind of backwards and forward and it’s really a safe space for people to speak about what they have done. If you are too harsh, people will not share what they have done and then that misses the whole point. I think it’s really important to have supportive feedback there and that it really comes across that you try to help them. Sometimes there’s also really good discussions about pros and cons and one of the points you can make there is that it never has any kind of perfect data visualization. There’s always something that is imperfect. Just give them time constraints and that’s good. And if people understand these kinds of different trade-offs, that’s a great learning experience. Then I talk about Gestalt Principles. So I show pictures to visualize these and better focus, really on only a few of the many Gestalt principles because I think going through seven, eight or nine depending on how many you want to show. It’s quite a lot of things. And then I asked them to redo their schedules. So depending on how much time you have for the training. Usually, it needs to be between 90 minutes and two hours. You can have a couple of different sketching Parts in it. If it’s 90 minutes. It’s probably more like two sketches. If you have 120 minutes, you can probably do three or four. That also depends on how many participants you have. So if you have 20 participants and you want to go through lots of these different data visualizations, of course then it takes much longer. If you have only six participants things can go quicker and of course, it depends on how interactive people are. After another sketch, I talked about eliminating clutter and there I used a really nice example from Cole, who knows Batman Affleck, about when people start shopping for the holiday season in the US and its really cluttered 3D bar chart, lots of grid lines in it and things like this. And I show them the first  really cluttered bar chart and ask them. What would you do to eliminate clutter here?

What can you take away from speaking about really good design is the principle of elimination. It’s about taking as much away as possible, so that the function is still there. But only those parts are still in there that are really critical. And usually they come up with all kinds of different good suggestions. Then I show how step-by-step she goes through this example and take away different elements of clutter like the gridlines, also tick marks.

 One of the things is that she’s really great in using gray to have different layers. Then I can speak about the power of gray on a white or black background and how you can have things that are less important. Put more into the background so that they’re not dominant things like axes, labels, tick marks and one of the things in this bar chart which says these different numbers for different parts in it. And it completely takes away the y-axis which very often creates a Wow effect, especially for statisticians, that it’s so used to show the y-axis and that’s really an interesting exercise. And very often some things that people remember and I think especially ask them first. What would you eliminate to create senseless expectation? Is that right? Or is there something that I’ve missed that can also be taken away and that’s usually something like that. Then I do potentially another sketch and then depending on the audience. I will talk about variability for statisticians, in the healthcare sector. I will talk about how you can potentially show individual patients. How can you show random arrows, how can you show confidence intervals. How you can bring in distributions, all these different things. I think this is really important, especially for statisticians, to be able to communicate with non-statistics. 

People that are, let’s say more beginners in terms of data visualization. It’s already a very complex topic. At the end of the training, I always close the loop, at the beginning, I set the expectation and then I ask about what you have learned and then I can see whether the expectations are met and that usually is a great way for them to reflect on the training to reiterate what they have learned and I really like this format in the beginning tell me about what they will learn, then having them learn it and then reflecting about what they have learned. In that way additionally will help them to remember things to reflect on that 90 minutes or 120 minutes well spent and that creates some kind of feeling of accomplishment. I think it is really important for these people to go and tell others about your training that these people refer you to maybe their supervisors to say. Hey, I had this great training. I think our whole team should get it or and these kinds of things really help to spread the word and that you also get really nice testimonials about your training that you can then further use to promote the training thereafter. And then I just kind of end with. Here’s a couple of references that you can read about and learn about more. Maybe there are additional things that you’re doing within your company, within your organization, in terms of data visualization training that you can use there and then do follow-ups. There are some surveys. I think that by now you will have pretty standards that you ask for feedback about it. And that will also help you to improve the training, fine-tune things over and over so that you get better and better. And lots of the features that I just talked about. I learned about this kind of follow-up questionnaire. So, that is how I teach people in terms of data visualization. Go to our home page and they will see all these different steps. In terms of setting expectations, having a case study, telling stories, engaging with the people, asking them to sketch things with pen and paper and then introducing Gestalt principles, eliminating clutter. I am also the preattentive attribute I talked about. And how I close the loop in terms of asking what they have learned.

 Have fun doing your own visualization training. And if you want to learn more about data visualizations, there is actually a whole side of the effective statistician set that is dedicated to resource training. All about data visualization. Have a look there, and use these for your data visualization training as well.  And I’m pretty sure you’re also learned a lot about certain things. That you can include in any other training be it about Base about sample size about descriptive statistics whatsoever. I think it’s really important that we as statisticians are really good teachers. So stay tuned for more. 

This show was created in association with PSI. Thanks to Reine who helps us with the show in the background and thank you for listening. Head over to where you will find lots about visualization. Lots of training material, lots of different things that you can use to improve your data visualization skills and also materials that you can then use to help others in this aspect as well. Reach your potential, Lead great science, and serve patients, Just Be an Effective statistician. 

Never miss an episode!

Join thousends of your peers and subscribe to get our latest updates by email!

Get the shownotes of our podcast episodes plus tips and tricks to increase your impact at work to boost your career!

We won't send you spam. Unsubscribe at any time. Powered by ConvertKit