Home / Healthcare & Medicine / Coding & Programming / Bioinformatics for Biologists: An Introduction to Linux, Bash Scripting, and R / Interview with Andy Kirk Part 2 – Visualisation: Art to Empower Data
Interview with Andy Kirk Part 2 – Visualisation: Art to Empower Data
Visualisation: art to empower Data Andy Kirk
We are facing today a very massive amount of continuously released data. And the problem is that, with this increased amount of big data, comes a flood of publications with visually unreadable, and sometimes really uninterpretable results. I know that from the big networks. Whether we know that an appropriate data visualisation lies in there really at the heart of impactful publishing. So what would be your advice to others to publish their findings in an appealing manner to allow the correct output reading, the understanding, and reuse by the scientific community. So are they kind of requirements in terms of aesthetics level, for example, and but that’s part of what art you were talking about.
Yeah, and I think this is where we do perhaps advise starting with the notion of simplifying. And I think one of the things that’s inevitable and it’s understandable, but one of the traps that a lot of researchers and academics will get into is, thinking that the only people who will be reading their work are experts with the exact same knowledge, and the exact same interest in a subject. And that’s very rarely the case. Even if you are sharing things with scientists, not everybody has the immediacy to understand what’s being portrayed. So the first thing is, start off simplifying things. Remove redundancies. Remove the things are overly technical that can be explained in a more accessible fashion.
Now, you talked about aesthetics and appeal, and I did mention that as a key principle that you should aspire towards. But I also think in terms of the journey, that some go through in developing more refined skills in this sense. If you start off by simplifying things, if you start off by trying to just clarify things, a certain elegance will emerge from that. Because people will experience it in a way that is accessible, and that in itself becomes elegant. Now, as you become more confident, and through repeated exposure, repeated practise, and by looking beyond your domain, and seeing what people in the media are doing, people in sports analytics, in retail.
Look beyond your domain And even if it’s not about the same topic that you are visualising, it may still be about how someone shows change over time, or how two categories compare, or how two variables seem to call some relationship. Everyone is doing the same analysis. It’s just about different topics. So don’t just think that your world is unique in terms of visual treatment. It won’t be. The meaning will be. But look elsewhere. Be inspired, and just gradually clarify, simplify, and then you might start become more ambitious, and then you might start having more confidence and more flair, to seek more appealing aesthetics. But don’t try and reach that overnight, instantly, in a single step.
See it is a journey of improvement, of continuous improvement. But step by step you will find success. Because people will be able to see what it is you’re saying. Yeah, I really feel that this is really important in everyday life. So even if you’re looking for a job, or I don’t know, presenting your last project in your enterprise, or whatever. I mean, even in the life of a scientist, there are moments in life where you have to present data in an adjusted way, data that you spent hours or years to release.
And I read somewhere that it might take you about 80% of your time just cleaning the data before visualising it, in order not to overload and lose that message. So what would be your advice to a person willing to present– let’s imagine a three year long project to scientists and or non scientists, not in 3 minutes as we usually see it now, but explain it in three main figures. Would you rather recommend an out-of-the-box approach, as you were saying. Or also, because it is important for them to understand and rapidly get that message, to do it an innovative way, but stick to known graphics, and something that they are used to see, but presented in a different way.
Yeah there’s a few things about that. First of all, it’s very easy for us to look around at all the wonderful new, innovative ideas, and think that there’s no role for the bar chart or for the line chart anymore. Because they’re old charts. We need to move on. We still need those. Those are the workhorses of the field. There’s a time and place for every chart. But crucially, every chart answers a data question. So one of the key things that we always need to establish, before we get excited about the ways that we might present our data, is to determine what it is we are trying to answer through an individual chart panel, or through a sequence, like you mentioned.
So when you are a researcher, and you spent three years, and you are so close to this data. You need to find a way to step away, to take that breather, to think clearly about, OK, what is it that someone looking at this work will be able to get an answer to. And write it down in language. What are the curiosities that you’re trying to answer? Because there’ll be a bunch of chart types, or even just a chart type in itself that will be the best method to convey an answer to that question visually. So questions are crucial. Questions are everything. In fact, questions come before data. Because we collect data in response to a question.
So get into that habit. Get into that behaviour. Write down questions. But I think secondly, don’t be precious. I find that for good reason, a lot of scientists that I’ve worked with feel that every single data point is important. And it may be that every data point is valuable, of course, but there has to be a hierarchy. There must be some things that are more relevant, more important than others, for the constraints of what you mentioned there, which is a very short, brief engagement, an audience may have to look at something, and to understand something. So don’t treat all data points equally. Some need to be suppressed. Some need to be elevated.
And find that editorial hierarchy that will allow you to focus on the key things, again, given the constraints that would exist. There’ll be plenty of other opportunities to share all these fine details, but not in every situation to every type of audience. So sometimes we get a little bit doubtful about the notion of editorial thinking, thinking that somehow it’s massaging results, or hiding things. No, it’s about editing. It’s about choosing what to include, and what to exclude. And we do that in everything. We do that in conversations. We do that in written form. We do that in text messages, tweets. We choose to say things and how we say them.
So that perspective of editorial thinking and question writing will help people to navigate through those constraints. Yeah, I’m really happy that you addressed this question this way. Because we are seeing in that particular research field, you’re talking about it in a very large sense. And I very much love that, because we’re facing these challenges every day, even in a single Twitter message. But really addressing it from also a researcher perspective, is really important. Because we see students sometimes, or even concerned researchers, presenting so much data at the same time, that you really lose that important message.
So one last thing I wanted to discuss is, I read a sentence that Henry Ford was saying about this topic of coming together is the beginning, he said. Keeping together is progress, and working together is success. And in that particular field, I thought that it showed the importance of a dynamic approach to data visualisation, and that should be driven by the fact that before actually you’re able to sell a bin as we say, you really have to know more about the person who is going to buy it. So I know we touched that subject a little bit in the previous question, but I really want to focus a little bit more on that to finish on this important topic.
Do you think that when we have to visually present our data, we should present it not in the way we see it, but in the way the audience will see it, and continuously keep adapting yourself to that particular audience. The same data can be presented in ten different ways to ten different people. Correct. Now, it’s not always practical to be able to share in ten different ways. And I think the first thing to acknowledge is, and again, this touches on the notion of editorial thinking. Because sometimes, you just have to decide that you can’t serve everyone’s appetites. There’ll be things that this person needs to know, and this person wants to know over here. You may have one chance.
So who are you going to feed? So you have to think about this sense of prioritisation. You have to think about the sense of, OK, as long as it suits that person’s needs, this person over here may still find interesting contents, but if I try and do both, I will actually potentially do neither. So that judgement of who you will not feed, I think is very, very important in this case. But yeah, I mean, the audience is something that’s very easy for us in this world to say, design for your audience. Design for your audience. What it means is quite subtle. Because it is about the characteristics, the capabilities, the confidence. It’s about the where.
Is it in a journal article, very small size graphic? Is it on the mobile, or a tablet, on the move? Is it a huge display in a poster, where there is a presenter alongside the poster, and you can have a conversation about the piece. So there’s lots of dynamics about the situation that someone would encounter a piece. But I think in terms of the idea there of collaborator think and teamwork, I think this is really important. Because sometimes, we try and do everything ourselves. And even if we can just introduce a second person’s mindset, ask a colleague, ask a critical friend about, does this work for you? Does this help you understand this?
Feedback is really something we reluctantly seek out, but we should. We should ask people to test things, run things past people. And even if that is not a perfectly representative persona of the intended audience, it’ll be better than none. And sometimes, we ourselves as creators, get too close to our work. And we lose sight of what it is we’re trying to say. So collaborating with others, getting other people to check things and test things, at the very least, is something that is going to give you the best chance of having the biggest impact with what it is you produce. Thank you so much for all these advices, Andy. You gave us a lot already.
Thank you very much for sharing this with us. Any last advice or recommendation that you would see important to discuss here. Yeah, I think what’s important is this is a subject that is attainable for anyone. We can all do better in how we communicate data visually. We do not have to be the most advanced technologist, the most brilliantly talented designer. It’s all about caring about the communication, what it is we want to say, how we want to say it, and who it’s for. So gradually, you will get there. You’ll find a way to improve what you’re doing. It’s not something that’s overnight, necessarily. But you will find a way forward.
So Andy, I am sincerely grateful for the time you dedicated to share all your experience with us. And I wish you all the success for the future of your activities. Thank you. Thank you.
This is the second part of the interview with Andy Kirk, data visualisation consultant, training provider, teacher, author, speaker, researcher and editor of the award-winning website called visualisingdata.com
Watch Fatma in conversation with Andy.
In the particular field of Data Science, we are facing a massive amount of data that is being continuously released. With this increased amount of big data, comes a flood of publications with visually unreadable and uninterpretable results.
Andy explains what his advice to authors is, so that they can publish their findings in an appealing manner and to allow understanding and reuse by the scientific community.
What is your experience in presenting data for publishing or for any other purpose?
What is your opinion on how important aesthetics is when trying to visualise data/results?
This article is from the online course:
Bioinformatics for Biologists: An Introduction to Linux, Bash Scripting, and R
Created by
This article is from the free online
Bioinformatics for Biologists: An Introduction to Linux, Bash Scripting, and R
Reach your personal and professional goals
Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.
Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.
