Interview by Aurelien Chu, Chief Operations Officer at Eskwelabs.
In this rapidly changing world, Eskwelabs introduces "Skills of the Future," a content category that features interviews and conversations with leading experts in technology, data, education, and business. We believe it is critical to keep up with the latest trends and developments in these areas and understand the skills that will be in high demand in the future of work years.
Communication has always been very key to any organization, whether you are in a startup, a corporation, or in academia. The way that you communicate to people has always been critical.
Ken: We have to accept that the burden of the communication is always on the communicator, not the audience . If the person on the receiving side doesn't understand what you are trying to do, then the responsibility is on you to make yourself understood. That is the role of data storytelling.”
Ken: I think the most common misconception that I see about data storytelling is when people think it’s just about graphs. Good data storytelling is fundamentally good storytelling. I have seen many analysts, especially in my background , who prioritised visuals over insights. Sometimes they show a graph on a slide - that’s it, just a graph with no context. That’s one of the things we can’t do in business. When I stepped into this role and I started making presentations, my mentor and boss, Lyn, would always ask me, ‘what do you want me to get from this slide? what’s your storyboard?” Since then, I always relay this to my teams and my students, you have to focus on the insight. What is going to move the business are the insights.
Ken: Traffic light visualization color schemes are fantastic, because they connect to the way we already think about certain colors. I think default tools has done a great disservice to many by providing default color schemes, because for many analysts, we aren’t actually putting much thought into the colors and their meanings, but that is a huge part of storytelling.
Ken: Honestly anyone who is really willing to learn and to take time to immerse themselves can learn this. At the beginning of my career I was really just into modelling, but as I grew into the role, it became more clear to me that this is an add-on to my skills, and it’s really important. The more you practice, the better you get at it. I started off thinking that storytelling was just making graphs, and my focus was on complex modelling, the communication was an afterthought for me. This is one of the things that has changed for me, as this is now such a core part of what I do and what I teach. I believe that a lot of passion can come from being good at something. As your competency develops, so can your enjoyment of that skill. As you put in the hours and get better at data storytelling, get more practice at it, you can really start to appreciate it.
Ken: In my case doing visualization was not really something that I intended to do at the beginning, but over time I grew to enjoy it, develop my own principles, and create a structure for personal learning in this field. I subscribe to the 10,000 hours’ hypothesis of Malcolm Gladwell.”
Aurelien: I agree, I think data storytelling, from visualization to narratives, can seem at first to be a minor aspect of graph selection and frankly something most people don’t spend enough time thinking about it, but it’s as you get more practice with it that you start to really appreciate the nuance, the value, and the ways in which data storytelling can be both a technically rigorous and an intensively creative field that really mixes the analytic and the artistic or performative side of a human being. At the same time, it really brings you closer to your analysis as well, you can’t do data storytelling without really understanding the insights that you’re bringing, and the insights are the core of the critical thinking that goes beyond building models and cleaning data.
Ken: I think that tools will change, but principles will not. I’ve already seen that, because as long as I’ve been working in this field, the tools have been changing but the principles have not. Over time, the tools are allowing us to do more and more advanced visualization, from 2d analysis to 3d analysis, to GIS visualizations, to interactive visuals, the tools just keep getting more powerful and easy to access.
Ken: ChatGPT is an example of this. On a recent project I wanted to combine two different types of visualizations for a Python-based analysis, to create a kind of combination visualization. I used ChatGPT to generate the code for me to do that, which it successfully did, and that allowed me insert exactly the visualization I was looking for into my project. I’m seeing a lot of that happening, especially as we’re combining different tools, PowerBI into Jupyter Notebooks as an example, and AI tools like ChatGPT can help make those workflows easier. I code dashboards now that automatically generate daily insights for the user.
Aurelien: As you point out, end-to-end projects are multi-tool workflows, which have traditionally been difficult because you need to understand every single tool along the way, as well as how to connect them to each. And that’s just the knowledge of the tools - the really indispensable knowledge is the domain expertise and analysis side. Your background is in economics, not computer science, but you still need to code every day to work with data analytics tools. Maybe ChatGPT can help make those multi-tool workflows easier by providing some of the glue that makes it easier to integrate each tool into your project, such as when you’re trying to figure out how to make a specific highly customized combo visualization in a new tool.
Ken: Even with all these technologies, the principles haven’t changed. I still have to plan out the visualizations, I still have to understand the why, I still have to identify which insights are the most important, I still have to figure out how to adapt the storytelling to the audience.
Aurelien: Maybe better tools can free us to spend more time on those core principles.
Ken: So one project that comes to mind for me was a project I did at work on queue analysis. Queue analysis is about understanding how many customers go through a drive through, and how many cars and customers we can service. How can we optimize it? In the food industry, this kind of capacity planning is very important.
In the beginning, the model I worked on was very complex, it was quite satisfying to build this complex statistical model. We were very happy with the results, it was a very good model. However, we had to tackle the challenge of how to communicate the results to the different teams that would actually be able to use it to improve our processes.
Ken: So a key thing that we did is that we came up with simple, catchy “rules of thumb” based on what we could learn from the model. For example, “for every 10 sales, you should have this many customers left in the queue” or “this many cars left in the queue”. This was a way that simplified a pretty difficult analysis into simple ratios and ranges that designers could work with, so that they could then figure out how to tweak our systems to accommodate this. That’s a lesson I want to share with others - you can have a great model but at the end of the day the way to get it into the real world and used by others is to find those simple, catchy ratios or rules of thumb that can be used by others.
Ken: Another project I really enjoyed was the carbon emissions analysis project that is carried out in the Eskwelabs Data Analytics Bootcamp. It’s just a lot of fun to see the different ways students come up with to break down a very complex topic with a lot of data and possible insights, and how they build a story out of their analysis. My favorite example was a group of students who built these great visualizations of how much CO2 is linked to activity for a single cell phone in Singapore - that kind of storytelling was simple but powerful. I still remember it and it may have been up to a year ago!
Ken: I am looking forward to getting more involved in a more strategic role. I enjoy the day to day part of my work in terms of investigating specific metrics, but I want to move towards the bigger picture, in terms of answering questions such as:
Ken: This is where my career is heading right now as I spend more and more time developing and operationalizing strategies for businesses, growing from my starting point focused on analyzing operational metrics.
Ken: My current topic that I am currently reading a lot about is change management. I am talking to people about where do you focus the energy of the team to prepare for the future without compromising what you already have now.
Ken: One book that I would recommend is W. H. Kim’s Blue Ocean Strategy, which I already use in some of my teaching at Eskwelabs for business intelligence analysts. The Decision Book: Fifty Models for Strategic Thinking is another one that I highly recommend. These are some of the resources that I’ve found have helped me as I spend more time thinking about and communicating about business strategies to different stakeholders, and developing meta-models for myself to break down more complex business questions.
To learn more about our Learning Sprints and how they can benefit your institution, we encourage you to download our Sprint Catalogue. Together, we can bridge the skills gap and prepare future generations for the changing nature of work.