Data Visualization, AI, Technology, and the Future Beyond Visuals with Paolo Ciuccarelli

Data Visualization, AI, Technology, and the Future Beyond Visuals with Paolo Ciuccarelli

Show Video

That's why designers found a place here. That's why designers are now well positioned to work in the field of data is precisely because we are able to understand people and problems and context. Much better than other disciplines, but just because we are trained on that and other disciplines are not, so it's not a matter of being better or it's just be trained on that and we spend a lot of time in training students in user studies, user research and understanding the context and history and culture and all of that helps you understanding why a certain person may react in a certain way and another person may react in a totally different way in front of the same visualization.

Hi, I am Paolo Ciuccarelli, and you're listening to GGUTTALKS, double G, U, double T. Hi everyone, I'm Maria and welcome to GGUTTALKS, a podcast brought to you by GGUTT, I enjoy adding value and helping wherever I can, widening my spectrum of thoughts even if it can sometimes challenge the mainstream. All I ask is to rate the show, leave a review and share it. It's a fantastic way to help other podcast explorers discover our content. Now let's get started.

Paolo Ciucarelli is a communication designer and researcher, professor of design, founder of the Density Design Research Lab, an award winning laboratory for data visualization and information design at Politecnico di Milano as well as founding director of the Center for Design at Northeastern University. He's committed to making complex data accessible and engaging, especially for those not traditionally versed in these fields, with a fundamental starting point in education. Paolo's contributions to the field have not gone unnoticed, with numerous award winning publications and speaking engagements worldwide. We explore Pablo's approach to design and data within context, complexity, interpretation, and accessibility, AI and technology, and the future beyond visuals.

A special shout out to Gianluca Brugnoli who made the intro. So, here we go. Paolo, I'm, really excited we got this big intro from Gianluca Brugnoli who came on a couple of occasions to GGUTTALKS.

Thanks Gianluca. I'm particularly excited about that because, it's a great topic we're gonna delve into. How, how did you get into the data visualization, information design space? What was the trigger? Yeah, it's a it's a long journey actually, because I didn't, didn't come at first. It was like, you know, step by step. Everything started with, uh, with, with a concept that I learned about actually long time ago when I was studying architecture, so my background is architecture. And while studying architecture, that's a story that I often tell, but I really got, fascinated by the idea of complexity, and it's a concept that I didn't know about, uh, you know, yes, as a term, but not really what that means.

And it's something that I discovered in the course, Politecnico di Milano when I was studying architecture, and it really changed the, the way I was perceiving, you know, things around because this idea that everything, especially social and natural phenomena are complex by nature, and it's something that you can't escape wasn't really, something I had in mind at that time. But then, since then, I started to see things differently with this idea of complexity in mind. And it's something that later I had a chance to bring as a topic for the students once, you know, I was, I became part of the School of Design when it was born at Politecnico di Milano. And then when communication design became available as a program, and I had, a studio based course I decided to challenge students with this idea of complexity and in that case specifically with trying to represent to make this complexity visible and understandable to other people as I understood complexity when I started to read books and wasn't easy.

I had to read very, complex books, you know, complexity brings complexity and, and it was difficult, you know, I had, uh, was a struggle. I wasn't particularly happy at that time to read those books, but there was something that I wanted to bring to students. And so we started to play with, I would say more diagrams and systems representations than data visualization. I didn't know anything particularly about data at that time, but step by step through the work also of the students,, we discovered because I was already building, you know, the the core of the research lab that eventually became density design.

So at that time, you know, I, I started to look at data and to discover data with the students as something that could make those diagrams more, let's say, alive, you get data, you, you feed data into the diagram and then it becomes like alive. It's like having a skeletal, a skeleton and then putting on the skeleton, some flesh and blood, and then it becomes real, becomes animated and alive. So that's why I got into data visualization, not because of data, because of this idea of this mission. There was has been always the mission of density design the research lab to make complexity we say visible, accessible, understandable and manageable Even though we know that it's not manageable entirely, at least, you know, you can orient, uh, complex phenomena. That's the best you can aim to. It's like, you know, renting that behavior, in a direction that I think it's appropriate, but that's, that's how I came to that.

And so data, I think that's, that has been an advantage. So the fact that I wasn't starting with data, so data was just a means to something as visualization is also a means. to something. So, and the goal is something else is to make people understand correctly, you know, in the best possible way what's going on and make them able to make decisions, you know, within that complexity, you know, not reducing too much because then what happens is that decisions are not effective as they should be.

So yeah, that's how it came to, to, to data visualization, not through data visualization per se, that existed, you know, when I was, I didn't know about it. And then eventually we got in contact with the community of people, especially computer scientists at that time working in data visualization. But we were like, you know, strangers, strange piece and that space where designers were not really, uh, present in that. And I think we contribute in a way to open a space for design for a different approach to data that was not data driven, but more like human centered. We say in, in design. Yeah.

So I, I stop because I could, I could, go on for a long time, but that's how I, I came to that. Um, no, thank you for this. I think I filled up almost, uh, three quarters of my page, there's a lot to unpack because you're touching on, on so many points here. But first do you have this feeling that data visualization is more hype today, especially with broad access to generate generative AI. It was a hype a few years ago. It's been a hype.

Now I think it's less of a hype, also because other hypes, as it usually happens came. And so this got a little bit, but it's also normal, something that, uh, wasn't, uh, you know, diffused as probably should have been. And so that's the hype. So people realized that was something interesting. Useful. And they started to ask for that.

Um, sometimes even not knowing exactly what was the purpose and how it could help. But on a certain point like you reach a certain level of maturity. And I think we are at that level of maturity that is now kind of taken in a way for granted. So people know that it's useful and should be used.

Um, and so it's good in a way because you don't, you get more, mature questions, you know, and requests so people are more aware, not enough, there is a lot of stuff to learn outside for in terms of what data visualization really is and how can it help, especially business, um, people, but, it's, it's more mature the field, so that's good. It's not a hype. So the demand in the market is a little less, sparkling, but, it's, it's fine. Now it's more challenging if you are in contact with business doing this and as I am. And so it's, uh, you have to fight a little more for clients, but it's, it's, you get more mature requests. Yeah.

And on this note, actually, can you expand a bit on your view? And if there is a difference because on how to approach things, if you want to, let's say, work with a data visualization designer or a data visualization engineer, because there is sort of an overlap and what I've seen is that, let's say engineers, let's say science driven people want data visualization engineers and not the word design And designers would want to collaborate more with data visualization designers because they can speak the same language but there is a big overlap from what I read also in conversations. Can you elaborate on on your view and experience on this? Yeah, I mean, to me, what is the purpose? So what are you? What are you searching for? So it's not because you are a designer or an engineer that you should look at engineers or designers is what you want to have. So, for certain purposes, the engineering analytical scientific approach, it's perfect and it works very well for other purposes of visualization, yeah. It doesn't, and that's where design can help going a little bit beyond the standard, creating more context for the data that you visualize engaging on expert users for example, that's why you need designers, whether you are an engineer or a designer yourself, you can't do that only, approaching the issue as a technical problem. So you need to have a different approach and that's why design comes into play.

So it's not, the. Who you are is what you, what you need and what is the purpose of the visualization, the visual interface that you are, that you're going to create, given that, then you assemble a team. I have to say that it's often made of both, you know, the designers. Yeah. And, uh, or the person that you, that's what we search for example, when we hire people.

We search for hybrid profiles, like, uh, either is a designer with a strong passion for hands on and coding, or is an engineer with a good sense of aesthetics and human centered research and very, Uh, close to people, needs. And so that's it may be, but it's the ideal, as you said, is that you have this kind of double identity in a way as a person, as a, as a designer or engineer or in a team, for example, that's more often the case. So it's not uncommon nowadays to have, for example, joint appointment as a colleague of mine at Northeastern University is having, for example, across, uh, these two colleges, actually design and computer science, or we have at the visual agency, for example, we have people that have both skills or both, you know, identities and can help. Yeah, they specialize usually in one or the other, but I can do both actually, they understand both. So these hybrid profiles are very not easy to find, as you say it, but, but very very, very useful and very valuable, in terms of, what is produced at the end.

Actually, this is why I was asking if they're hard to find. I think in my mind, I was told because I had something kind of specific to ask you in that, space because The other thing is, how do you educate customers on that one? Before educating customers, you need to educate people that are able to do that. So, and honestly, one of the reasons why in 2019 I decided to move to Northeastern University in Boston is because they have, they had, and they have still, a master program in information design and data visualization that doesn't really exist in many other universities. Not at that scale, not with those kind of faculty, so in Politecnico di Milano, density design is, I think, still the only place where you can do data visualization and information design so there is no other course, no other research group doing that.

And I think it's a limit, you know, given the importance of data and the importance of making sense of data, the fact that a big university, public university is very relevant doesn't have, for example, a program dedicated to that, it might be a challenge for the future, but it's, that needs to be, there should be more of that. So, one of the reasons why it's difficult to find those people, because there is no, there are few schools where you can train them, it's growing, slowly, it's growing a lot online so there are courses with dozens of thousands of, of people attending, by, some of the most popular designers, information designers and data visualization people that they do these courses. And this usually are very well attended because, and there is a demand also for training, professional training. So people want to learn more about it. And it's needed because then, you can find a job if you do that.

But Yeah, there is not enough, certainly not in university programs. So there are other organizations and entities trying to fill the gap and cover this space that is indeed in need. So you answered my question because I heard of Density Labs many years ago in Milan, obviously, and, and you just said that now I know why you moved because I was going to ask this. Have you ever thought of just starting your own labs, I don't know, like a licensing or an online school or something like that to spread, the word more because you're talking about hybrid figures. Yes, in a way, yes. And we did it, for example Politecnico has a consortium that is for a post graduate education and, uh, and we did a couple of years, a course that was mostly for professional people and it went very well.

It was full, fully booked and well attended, but then we had, that was just a matter of time, we couldn't, but this year, it's restarting. So PoliDesign, that is a consortium of Politecnico, di Milano will do that again. And, in, Northeastern University, I didn't feel that need because there is a, an entire program. And so, you know, there is a lot to do in, in curricular, education. And also probably didn't have the time to think about that. I moved four years ago, so.

Still kind of a new context, but, and, and then thinking about the third entity, and connected with that is the visual agency. We, we, we did in the past to some training sessions, for example, just for a company. So you get the managers, line managers, and then you do that or.

Now it goes under the label of, let's say, data culture. So there are a lot of data culture initiatives within companies. And in, in many cases, I would say in the good cases, then you have also data visualization as part of it, because you can't have, let's say, data culture without talking about visualization.

That is one of the, I would say. stronger cultural components of the data processes. And so, so, yeah, I, I thought to that and I did something in those hybrid spaces, meaning that is in between, the academic context and the professional context, but you can do that also as part of the university. And that's something very important, even in a public university as Politecnico, you have this consortium that only does that, you know, education is transfer the knowledge to, to people that work professionally and, the same at Northeastern University and then, well, the visual agencies, of course, is more business oriented, but they use also, other faculty and professors to, to do this training. So it's.

I think it's this positive mix of, of, uh, academic profiles and, and, business, business goals. You said, let's go back to data because I think now,, you kind of mentioned it before and now, and here, I mean, data is everywhere but it can mean or have different connotations to different people. So before we delve a little bit deeper just to set the scene, that's a big like, what is data? That's funny because I, I'm running a little, module of a course that is about strategic design, so something a little broader. And my four lectures are about data and creativity, and, and of course. We started by, reflecting on what data is, and so that's how I usually start. What I know is especially what data is not, and this is, something that, isn't still obvious, like, you know, what data is not.

That data is not, for example, that objective neutral entity that we were used to think of in the past is not obvious that is a, is a construction. It's a human made, object that is eventually constructed in a certain way and is, uh, framed from very different perspectives you know, technologically, economically, philosophically, that's what we discussed in my lecture last, last week, for example, all these frames that data has that you should know, because otherwise you don't know exactly what that means, because a data point has been produced for a certain purpose by someone with certain technologies for certain reasons and with certain knowledge or ideas or what we call biases nowadays and nowadays it's more clear. Thanks I would say to big data first and then algorithms we discovered how humane are those technologies. And now we are bringing that back to data. And so, oh, wow.

So data was the same was exactly as humane as algorithms and, and, and, artificial intelligence. And, but we didn't think about that. So you had also terminology that I think was profoundly wrong. So people talking about, digging data or like extracting value, but the value is not there. I mean, it's something that you build intentionally, culturally, technically.

So it's something that I would say is an interpretation, the data itself. So you have a phenomenon, you want to try to understand and make decisions and act. You need an approximation because they can't, you know, deal with the phenomenon one to one real scale.

So you reduce it with the data, but you reduce it in a certain way. And that's your first interpretation of the phenomenon because you're a human being. So you can't be neutral.

And so, and then you visualize it and that's another level, a second level of interpretation. And it's even more effective in certain cases, or maybe not because data is the fact is all the interpretations that you do with data are invisible. transparent.

You don't see that people can't don't have the means with visual it's more understandable. So you can, you know, discover some intentions in certain ways of designing visualizations because you can, drift people in very opposite directions, even just changing the way you represent the data set. But you can guess, what was the intention behind if you master a little bit Visual languages, but with data, it's difficult if you don't know about data languages and data processes, it's very difficult. So you can trick and then, you know, data and you can even lie, as people say, you know, you can lie with statistics and data, but, and people will never realize that or very, very few people can realize that a little more with visualization, but in both cases, you, you can really interpret, and then just a last comment on data and visualization is that it doesn't end there. So because you produce your visualization and then maybe that's what you meant before with your question, then it all goes to the reader. I can have one visualization and put that visualization in front of two different people and get a totally different insight out of it because of the null prior knowledge of that person regarding the phenomenon because of the skills and the competencies that person has.

You can really trigger very different insights into different people with the same visualization or have the same visualization when one person would get the insight and the other one will not. But that doesn't, it's, you know, the visualization is never enough to get an insight. It's all, it's in between, I say, that space in between the visualization and the mind and the eyes of the reader.

That's where the insight happens, but it needs both. It's not there. It's in that space in between the reader, the analyst lay person and the visualization. Going back to what you were mentioning before, the context, right? There's always a context around what you're trying to visualize and convey, but yet you will have different understandings. And this goes, I mean, the prime example, which is not the nicest example in the world, but this is wars, right? You have two different, perspectives, two different stories, at least two, I would, and more.

And this is at a bigger scale, and I'm talking about, mass media kind of information where it's one or the other, and then, or one, but depending on the way you look at it, you would have your answer because you have your, preconceptions or whatever and then if you talk about health data and medical data, how can you, talk about both, like make a comparison. I'm just trying to see, like, if you have a project you could, uh, explain just to go to be a bit more tangible in the space of health data, you mean, or it doesn't matter in general where you would have different interpretation, maybe take the most complex project you were worked on and how you managed to communicate because you spoke of decision making, right? So it is. Sometimes these decisions can be costly, but they're just decisions because decisions need to be made. No, I mean, projects might be complex for very different reasons.

Some of them are complex for because of the technological challenge. , for example, in one project, in, in, Politecnico di Milano, on a certain point, we wanted to visualize all the students, you know, one by one as dots in the screen. And that time was challenging because they were very few frameworks that could allow 44, 000 to move dots to move smoothly on the screen, and that's something that, of course, then, you know, has technology advances, then it becomes less of a challenge. Another complexity another level of complexity for project comes from from the people that you involve in what's the user and how many stakeholders you have in that regard for example, for me, the more complex, the more difficult is always the let's say the lay person. So the non expert that's, that's always difficult because you have to recreate a little bit of context because they will not understand, the more you go far from the source, you know, what data has been produced, the more you have to add knowledge and context that the person doesn't have. That's to me brings additional complexity.

So the more you are far from the expertise in terms of data, data analysis, the topic, the domain. The more I need to add, and it's not easy to understand how and what to add. These are two different types of complexity. There are others, like, you know, the types of data that you're managing.

And when you talk, for example, about health data. That's really, you know, challenging. I remember recently we started a project that, you know, fortunately didn't come to an end, but because it was a startup and, you know, with startups, it's like you bet in a way on your client and, uh, but it was a very interesting challenge because they were performing some novel analysis on people and getting a result that you couldn't have through traditional or available analysis. And that, result was about something very important and very You know, like life changing event in the life of a person.

And so how do you communicate that result? Technically, there's a number. So very simple. How you communicate such an important data point to a person that after seeing this data will change probably forever, you know, his life. And so we'll see change to the life, but so that's something. And so we try always again to be As less reductive as possible.

So, for example, it was an index, but we immediately decided to unpack the index and try to explain the different components of that index in making that not looking like yes and no, because, because actually it wasn't really the case. And it's not that is sometimes it seems that you have a threshold, but it's not exactly. Combine these nuances, this, um.

This potential spaces of interpretation of a data. I think it's in that case was very important. And it's not there is no convention to do that. And you have really to work a lot on studying the user.

And but health care in general is a lot about that. Data is often represented in a very simplistic way. And it could be misleading. It could be scary. It could be you could create false expectations and really hurt people in a way that you shouldn't, shouldn't do. And even the hospital, there are ways of representing data that could be changed, even sound, I'm working nowadays a lot with data sonification that is translating data into sound and not just visual, but in most of the cases, it's just, alarms and then you know that something that's that is just meant to attract your attention.

But then it create maybe an anxiety where it's not needed. And in health care, that's important. It's definitely one of the most complex areas and, and where data might be really simple, but problem on how to how to convey the state and how to make people feeling right about that and understanding and not be misled. I think that's that's not easy at all. Here we're talking about, it's like information that you need to just give people what they need to know what, what does it mean, actually, and what it means is can be different things to different people in different contexts, and they can understand it with visuals but with sounds.

Some are like a trigger. You need to do this now. We need to know this now without looking just we're talking about experience here, right? As you were moving along or whatever, but the visual is something you will stop and look at and spend more time on before. And maybe the longer you spend time looking at something, the more confused you can get, even if, is this what you're trying to say somehow? I mean, it's, um, that's something that happens, but, usually if people feel that they can get confused, they don't stop and look, they just go away. So, but anyway, uh, I think the, I mean, this going back to, to healthcare, I think your question is really brings, the importance of, of understanding people and the context back.

That's why designers found a place here. That's why designers are now we're positioned to work in the field of data is precisely because we are able to understand people and problems and context much better than other disciplines, but just because we are trained on that and other disciplines are not so it's not a matter of, being better or it's just be trained on that and we spend a lot of time in training students in user studies, user research and and understanding the context and history and culture and all of that that helps you understanding why a certain person may react in a certain way and another person may react in a totally different way in front of the same visualization. So that's something that you learn and something that you apply and something that you spend a lot of time on. That's difficult to, to sell to companies especially, the part where you, you sit, you know, beside the user and then you observe and then you try to understand and then you try to wear the shoes and You design for them, something that makes sense for them in that context at that time and for the purpose that you both know. Yeah, I think that's, the hardest, sell actually, but the most important one because if this is not done right, then anything you do and all the investments that would happen after might be just a waste.

I mean, it doesn't mean that you need to do that at all the times, you know, there is literature. So there are studies done, that's what we do research, you know, we do studies and then we publish, and then, you know, that certain things work for certain types of people. So you, you have that, so you can standardize. That's not that you have to restart all the time, but if you are in front of a new problem in front of a new audience, In front of new type of user that is approaching this data for the first time. And I remember when first, uh, the first platform for, uh, trading online were available a long time ago. And what they did, you know, the banks, mostly they just opened the visualizations that were available to the professional traders, to the users that had no clue about, you know, and I, we, we did some status at the time.

And it was clear that, you know, the, the reaction of, of a, of a normal person, the owner of a bank account in front of a trend that is going very fast, high and high, you know, then you get excited and then it's all green is not, is then you start behaving in a way that maybe if you knew something about the market and stocks, you wouldn't have done the same thing, but the visual was. bringing you somewhere. You didn't have the knowledge to really understand that you just looking at the data. So that's that's something that you need to know. I mean, you need to understand what is the who is the others on the other side of the visualization. That's the most important.

Yeah, and, and not always designing for the experts expecting that everyone will understand it. If the problem is new, the context is new, the user is new, then you can't rely on standards and literature because it's new. Yeah. Same now for, you know, AI and all these new things that we have. But still with big data, we don't have the visual means to represent all the features that We as humans produced with big data. We produce big data.

It's a different type of material. It's like comparing, I don't know, you know, normal poly polymer with composite materials. You can't use the same technologies.

You can't use the same and we represent big data with the same charts that have been invented in the 19th century for a very simple statistical Tables about import export in in Europe, but they were conceived for, maybe a few rows and few columns and now you have big data lakes, like graph databases, and then you still, push squeeze that into dashboard with very simplistic visualization. Yeah. Maybe if you have an expert on the other side, they will infuse all these, their knowledge into this dashboard, and then we'll come up with an insight anyway, but if you bring this visualizations to a domain expert, a decision maker, a policymaker, for example, in the case of government, they may know about the topic, but they, they probably were not able to get all the complexity of the problems. Just. All the complexity of the data because big data are complex, and that's good. They are more likely to approximate the complexity of the phenomenon because they are more complex in nature.

But we need also more complex visualizations in nature than the ones that we are using, and that's not happening. So it's happening with data, and that's fine. That's super. But it's not happening with visualization. So I want to see big visualizations as we have big data and not the same visualization that we were using with, let's say, simple data set.

And that's slowly happening, very slowly. But if you see the market of, you know, visualization tools, they are based on the same paradigm that we were using, you know, hundreds of years ago, actually. The structure is the same.

Yes. Do you feel that with some of the new generative AI platforms, at least, um, testing things out and what would be faster? It's changing things. I mean, that's that's it's obvious nowadays, but it's happening in all fields and it will happen also in data visualization. So my impression is that before having those big visualizations I was, you know, claiming about in response to big data, we will probably have a different way of different relationship, totally different relationship with data. That is, um, Probably brings more text, you know, in the way you, you see data described, because that's where AI is very powerful.

It works very well, pretty well, much more, much better than the visuals and data visualization, especially there is no, yet no tool that I know about that is able to generate, autonomously, the data visualizations and of the complexity that we need. But if you compliment visuals and text, that's where things becomes interesting. And so then maybe you don't need to it's, it's like another route for getting what, what I was, what I would like to get to that is this level of complexity. There is not.

That is enough to give you a good idea of the phenomenon, so you, you probably will get that through a different way. That is like, maybe more narrative I would say, like, we, we call that in we are trying to develop that both on the research side and the business side. It's like augmented dashboards or augmented business intelligence that where you have text and pictures work together to convey an insight and in reaction to the user that you have on there. So it's, it's more like there is a dialogue that is established between the person on one that is using.

So this idea of understanding a little bit more the context and the user is in somehow could be embedded into the algorithms that combine in a, in a multi modal interface, um, meaning text, visual, perhaps sound, as I said, and narration that reflects the insights that you're searching for. So that's, you know, it's very vague because that's what we are researching on nowadays, but we're trying to develop some, some first prototypes. But it's working. So some of the first prototypes are about, uh, visuals that you produce as a normal, let's say, data visualization operator or information designer. And then the text around and let's say the interpretation of those charts is generated all by a generative AI working on text.

And so that's something that can be that can create a different kind of reporting, for example, or the visual agency created this tool that generates motion graphics based on data automatically. So you just create some rules and templates, then you feed data about, let's say, I don't know. a hundred of different countries and then the visualization adapts to this data and then generates a hundred of different motion graphics that have the same family feeling but that react to the data and then produces at the end automatically like different visualizations.

This, this would take personalization to another level. Actually, what you're saying, uh, I think it can be very abstract to some of the listeners. We're talking about. I'll try to wrap this up. I like the idea of augmented visualization actually, I was having a conversation yesterday with a colleague of mine, and we were talking about, Data visualizations in particular, but that's and and he was actually saying we should put some sort of motion to explain what this means, because it's complex.

And what you're saying here is having this with voice or or sound and visualizations and, um, Um, And even later on with AR VR, not later on right now, actually, where do you see this going? Because the goal is, even though what we might be discussing might sound very complex, the idea is to make it very simple and accessible digestible, understandable, understood and actionable by whoever is looking at this to make decisions. I'm a little skeptical about the word simple when it's about complex phenomena. So it's, it might be like a journey made of simple steps or an experience made of simple steps. But all in all, if you want to see complexity, you have to have a complex interface. I mean, then doesn't mean that it's complicated and you can't access it, but you have to provide ways to get into the complexity of the visualization that is reflecting the complexity of the phenomenon.

I don't want to sound too much, but for example, you know, you mentioned animation, you mentioned one way is just, unfolding that complexity through time. So you enter into a narration that starts from a certain point, and then it's like, you know, navigating the links and the nodes of that gigantic network that is is the phenomena. But you're guided through that choosing one potential, you know, Path out of the many. That's a choice. That's a design decision, you know, and then you build a narration that you drive people through that complexity might be a long, you know, it's simple because step by step, little by little, you get pieces of information but in the end, you need to have that level of complexity that is enough.

That is enough to to reflect. So it's a trade. It's a trade off.

Always. It's in a very tricky that that's the trade off that we are constantly managing as designers. It shouldn't be too reductive because otherwise the decision to make is not effective. But it shouldn't be complicated because otherwise you will not look at it at all.

And again, the decision will be your ggutt and not your data driven. So you have to bring again this I'm always talking about phenomena that are complex in nature. So, but, you know, complicated nature means also an organization like a business.

I'm fighting a lot with the typical way that you see people represented in companies, in organizations. So we still, in some cases use this organizational charts, you know, pyramids, or that's not what an organization is, especially if it's a big organization with a complex dynamic, social dynamics that that happens would happen within a big company. Innovative companies, for example, they, they have processes and relationships that are not reflected at all. So, yeah, you can continue using those, but you lose, maybe 80 percent of what's going on and and how people really behave within the company. It's a pity, it will make the work of an HR manager much more complicated and, and it will probably do make decisions losing a lot of the information that is floating around, following path and and links and connections that are not represented by those very simple. Yeah, it's an organizational chart is very simple to me.

It's too simple, especially for certain organizations. And, um, there is a trade off and that's where we work is everything represented. All the employees, all the relationships they have. All the information that they get and share, but maybe something in between that is not too complicated, but it's not the organizational chart that just says who reports to who. Basically it goes back to what does it mean? It doesn't mean anything.

It's like who's, who's labeled how, but then what else happens is missing. Yeah. If you want to see that, if your purpose is to see, you know, who's reporting to who officially and, who is in charge of who and who's, uh, responsible for whom, that's fine.

But if you want to understand, for example, who's the person that is driving that group and really pushing this organization in a certain direction, who's the person that I need to retain because he's absolutely essential for the, for this organization. You don't see that not, not through the organizational chart, not through the, some of the dashboards that I've seen and you have a lots of data within the company. There are, you know, many constraints because of privacy.

And that's great because we don't want to intrude, people's lives and make decisions that, we know how biases might be relevant, but there are ways to handle this data in a way that's more representative of complexity. Because then What you want to avoid is at the end, organizations are made of people. You want to make sure that you don't lose the good ones and that you, you get good ones from the outside.

And, or you want to know what you lose if someone, some of them goes out. It's not just that it's an empty cell. It's maybe a hub for an entire network of connections and, and dynamics and relationships that you, if you don't see them, you don't know them. So, so going back to what you were saying earlier, just, just before that, that led us to this, you kind of eluded to wayfinding somehow through the visualizations, because you were saying you need to take people, let's say, on a journey through a path to uncover and discover along the way. And I'm just trying to put this into a context of physical spaces, you know, wayfinding and so on. But also kind of zooming in, zooming out where it matters the most.

And this is part of the You know, hierarchy, when you want to represent something, you start with the general idea, I guess, and then you go a bit deeper. That could be like an approach. How, how long would one, I mean, here we're obviously talking about, I mean, this could be physical and digital, uh, visualizations in a certain way, or, um, data visualization in a certain sense, how, how fast or how slow should one I know that there's no yes and no answer to this question, but isn't the scope somehow to, Because if we're mixing this with business decisions. Isn't it like I look at it.

I get it. I know what to do next kind of versus I want to have a museum like experience. I see you laughing. What? No, I mean, these are kind of two very far. Concepts or potential experiences of data. Yes.

Again, there is something in the middle or that's I think the most important is that you have, you should have different interfaces for different purposes. So I'm not against, you know, that's something that I want to be clear on that so I'm not against the business intelligence dashboard that we have nowadays because they work again perfectly for certain types of decisions on certain types of levels. And then, as you said, if you if you want to explore or you don't have a clear question or you that's this more exploratory approach to to the data you have and to your organization, for example, that's something that cannot be done with the interfaces that you used to compare quickly to numbers and then say, Okay, I need to do that. So that's something that's possible. And there's a long tradition tools work pretty well, I don't want to make names, but there are companies that produce wonderful visualization, visual analytics tools and, and it's fine. So there is a market and they're growing and, there are reasons for that.

It's not all, you know, not all the decisions you make in, in a company, in an organization, are those kinds of decisions and those other types of decisions, they serves other types of interfaces, other types of processing the data, perhaps, and maybe other data that you're not thinking to because you just have those other kinds of interfaces. So visualization has also this role sometime if you're going to certain routes, then you discover that you may need or may want to have data that you didn't think were useful. So there are plenty of opportunities to produce data. Sometimes you don't do that because it costs money.

And if there is no reason you don't do that, but you don't see the reason because you never tried, you know, to make some scenarios, for example, and say, okay, what if, or maybe you didn't think freely to what do you want to see or what do you want to know? Because you were caged on the data you have. That's very happens very often in the company, but sometimes it's not just that you have to produce new data. Sometimes data exists. Outside open data, data owned by certain companies. And maybe if you don't start thinking broadly to what is that you want to see, what is the problem, what is, and then you look at which data you need, and then maybe some of them you have it. Some others are available outside, maybe free open access, and maybe some others don't exist and then you could create them.

We worked on some projects like that. That was fantastic. Because, you really can change the way people look at data, but just, trying to visualize, you know, making. So what is it you want to see? What is your problem? What is without thinking about data and what you have? Or the tools, the visualization tools that you have, because if you limit yourself to the charts that are available on Excel, depending on the version that you're using, because each year they keep adding some visual patterns, but it's not all that exist. And we can invent other ways of visualizing data if it's needed. I mean, there is a reason if there is a purpose, and then maybe we create the data that we need in order to get that picture.

That's something that is also possible. The more technology becomes, available and performative, we can create data now is much more common again, thanks to algorithms. Now this idea of creating data like synthetic data, for example, it's much more common that it could have been in the past. So it's something that makes sense nowadays, but probably could have made sense also in the past to have some sort of synthetic data to perform some You know, speculative analysis, and then maybe think about having that for real. But, visualization is not just an end product is sometimes it's a process that you go through in order to understand. And then.

Can lead to change your perspective on, on data and what do you need? It's never ending. I mean, because it, it, it fluctuates, right? Data changes comes and goes, and there's more or less you need and whatever, just going back to the concept. There's so much to unpack here. I mean, I finding, finding this hilarious that we're doing this podcast that is audio and video. This is on YouTube as well, but then we're showing absolutely nothing.

So just where it's all right. Yeah, I'm sorry. I wasn't prepared for that. Um, no, no, no worries. It's, it's audio primarily, right? So for, so we have to, we, we have to be for everyone, but I'm just, some concepts are so abstract that that's what I'm just trying to dig a bit deeper. So we can, uh, it's, it's for me.

I mean, I just find this fascinating, to be honest. I mean, it's one off one for me, people to communicate, right? And then it evolves as, as we learn more and we want more. There's so much to unpack in what you said, but you mentioned your GGUTT. So my question is, which is obviously the main question I asked everyone on this podcast, since I started, it's what's your relationship with your GGUTT? Because you mentioned decision, you said, it's not going to be a GGUTT decision, right? Your GGUTT in your work, but also in your life? Uh, I think it's a good relationship meaning that I I I made decision that I realized, we're not rational somehow so they were a little bit more instinctual. And that's what I see when I think of GGUTT, you know, that's, is this, um, instinctual, I would say, opportunism. That's positive, you know, meaning that I tend to follow people or projects or.

Thing that emerged and that I so I would say that I'm a more like, I don't know how to call that that approach. But, you know, following what is emerging rather than, you know, planning 5 10 years ahead. But I think that's also maybe more common for designers to have this very open approach. Open eyes, open ears, open mind, and then according to what emerges.

And then you look around, you know, you have your tools and methods to look around, to observe, and then you shape things. But whether it's an organization as a research lab or a research center, as the one that I founded at the Northeastern University, it was based on some things that were emerging in all the cases was not just, planning in abstract from, it's, it's a mix of them, but I think the gut prevails in a way on the rational plan more often than not might be a limit, but so far it worked. And so, yeah, that's in terms of, GGUTT feeling, so to say, and how I make decisions and I.

Because you, I mean, it's very difficult to, you never have the full picture and have like the all information you need in order to make decisions. You have to rely a little bit on that, in any case. So the more you don't, you don't visualize, uh, or go do some extreme data. I do, I do, I do, not, not, not data visualization, but I, that's how I started. I always loved diagrams and and schematic representations, and I was doing that for, you know, even for other people when there was a meeting or someone was writing a book.

They always asked me in the past, to do some some of these charts. And it was the time when data visualization wasn't really a thing. So it was like more diagrams trying to explain concepts. And when I have to, I did the same, for density design. I did the same for the Center for Design and Artists.

And there is somewhere a sketch. Like a diagram that is reflecting the idea and then it's a little bit changing. I have somewhere those diagrams.

That's, yeah, sketching , and diagrams were my way of, you know, it's like the bridge between GGUTT and planning and rationalizing. But it's still sketched, you know, it's not final, it's not precise, it's not accurate because it's not totally rational, it comes from GGUTT and So I love those kinds of representations and, yeah, that's a good way to, to go through. It's like you think when you draw and that's, I think, something that depending on the way you express what you think, then your ideas change. So is this the way, you approach any project? What's your process? Like, do you start with rough sketches, rough diagrams and then, and can you take us through like a typical project? I know there's no clear, like defined, right? It changes according to project, but what's like as an overall, if someone's listening and is like, okay, I have something around, data visualization, how can they start? First you start from the data.

I want to look at the data and as less structured as possible. So like, uh, we say raw by raw data is not a thing. That's something that I don't like.

But, anyway, so I want to go to the source and see what data you have and then start thinking about that. Maybe visually and reflect on what this potentially can become, and then by also by myself, you know, without, of course, people have ideas, but I want to have my ideas on the data. And then after that you sketch, that's always the next step sketching potential ways of representing not the data, but the phenomenon or the insight that the person is searching for the user. And then you start thinking about the technicalities. So which chart or is, is a chart that exists is a library that exists is, doesn't need to be code coded, from scratch , and then of course, mixing constraints that you have time resources, then you go on direction, different directions.

But first of all, I want to see what I would like to have if possible. And then see how we can get there. Because there are more ways and possibilities to do something that doesn't exist yet than we had in the past. So before discarding something that is, doesn't seem to be available. I try to sketch it a little at least, and then. Let's see what we can do.

And I assume, obviously, the first step is you try to understand the context, right? And who's going to be involved, as I mentioned before. Yeah, absolutely. So, you said you try to understand what do you want to Get out of it and going back to Simon Sinek, start with why does the, why come after, No, no, of course, but that's, yeah, I come, come to even before what we said earlier, so I was talking about the, let's say the process, after, you know, after you, yeah, yeah.

It's even after I understand the context and context and why. So what is the purpose in that? And that's purpose contains why to me. It's not just a goal.

The function is broader. So in the purpose for me that is the why. And then. After that, you start your process, sometimes you just stop at the why, because maybe it's not something that it's worth.

Exactly. So, yeah. No, I wanted to highlight this actually, so it's good like you, uh, you're mentioning just in case. So, so it's not kind of misinterpreted after in any case, because.

Obviously, as designers, also, you know, it starts with Y. And as you said, exactly like sometimes you can stop after the Y as well. Yeah, I think that is, um, a colleague of mine says designers nowadays before designing the thing that should reflect on what are the things that deserve to be designed because not everything The service will be designed or should be designed. So let's first reflect, let's say on the why, and then because, and same for data, you know, sometimes certain data shouldn't be visualized, shouldn't be represented. And it would be better not to show something, not because you have it, you need to see it, or you need to visualize it and make it available. So that's goes.

a little bit to the why. And I think it's also an area where we can say something as designers, because it's not a technical problem. It goes more like on a philosophical level, but that's where sometimes you can make the good decisions before even starting the process. Again, so much based on what you were saying.

So You mentioned data and obviously we can get there is data that you don't need because some in some contexts, some people can be just overwhelmed with lots of data and they don't know what to do with it. That's one thing and Going back to the raw data, you said it's something you don't like, so can you, can you expand on this? Simply because what we said at the beginning, so because data is not, is, uh, it's never raw, it's not something that pre exists. You create it, so it's not like, raw, raw food exists.

It's not you, it's something that you fish or you, it's, it's there and then you get it and then you cook. But then with data, you intentionally decide, to create data. So it's not something, yeah, for you might be given because it's something, but it's not given in principle in the first place. It's something that has been constructed by someone, and that's something that you need to know in order to understand what you have in your hands. When we train students, the product designers teach them how to look at wood and recognize what kind of wood it is and what it comes from and how you can process it and if it's good or not, and certain companies nowadays they don't work with certain type of woods because it's not good for the planet.

So you need to know that and you need to know what kind of processes data went through before you get, they get to your hands. So that's what I mean by not thinking as less as possible today. There's something that is wrong. It's not wrong.

Most of the cases is. Already cooked to a certain extent. I will try to visualize this if we talk about, generative AI platforms, right? You have pictures that, before training you have the pictures, same here for the text before training the models You have, the words and the texts going back to what you're saying, the raw material, right? Is this what you mean? Because if we get another, not generative AI stuff, we talk about wearables. I don't know, Apple watch, right? Let's say smartwatches or, Oura ring or whatever. We're collecting data as we are moving we don't have access to it, I guess, as, as raw data.

For me, raw data is like, like an Excel sheet file full of numbers, something like that. So can you, can you explain what, what you mean by that? And if this is making sense? Yeah, I mean, if you think, for example, to the Excel table That's not raw data to me. You know, it's something that has a structure already. So you have to make decisions on the columns and the rows. And you are structuring the data set in a certain way, intentionally, or maybe unconsciously, but you are doing that.

The moment you put data within a table, now we can, someone says that table is the first form of data visualization. So it's already a visualization, a representation of the data. So that's not raw data for me it's already an interpretation because you have to make decisions in order to fill those cells in a certain way. You can fill the cells in many different ways. So it's like a puzzle that is completely deconstructed. This is raw data for you, or not even that? No, not even . Because a puzzle is

already, you know, you know there is one way to put pieces together, so you can't really do much more than what was already. planned Legos, Lego Legos a little better. . Okay. a little better because there are, different options. Yeah. But then you can play with them in a different way with the puzzle is much more difficult.

So yeah. Lego is a little more raw than, than , puzzles. But again, it's more like that. I, I want people to see, recognize that this table is already driving you. In a certain direction. So if you put, certain data in columns or, you know, you can make decisions that really drive the perception of data in different directions with the table, not even with the visualization.

So that's we are in the kitchen already, so to say, so it's not so that the data is the one that is, for example, the most raw data can imagine, as you say, you have a sensor, the closer you get to the sensor, the closer you get to the raw data. So there is a sensor, but then itself, you need to know. Which sensor it is. What is the level of accuracy? How does it measure, detect, you know, the signal that is biological, natural, whatever, and then transform that into a digital entity. So how does that work? Nobody knows, but there are different ways to do that, of course, and might be done in different ways for different reasons. So that that moment the sensor works, that's you create the first data, but then How this goes to the surface of your iwatch, then there is another lots of decisions over there.

And then you see data like circles or whatever. And then this, this is another that's all cooked in other tons of people working on deciding

2024-05-05 19:07

Show Video

Other news