00:00:04: The Biorevolution podcast.
00:00:06: Your hosts.
00:00:07: Luise von Stechhoch.
00:00:08: And Andreas Hoechler.
00:00:11: Everything, everywhere.
00:00:13: all at once, mapping the bigger picture with macroscopes.
00:00:17: No, no, you heard right.
00:00:18: The show is titled that way.
00:00:21: We're not talking about telescopes.
00:00:22: We're not talking about macroscopes.
00:00:24: We're talking about macroscopes.
00:00:25: What that is, we have a very skilled guest today from Bloomington, Indiana, and we're very happy to produce this show here in Berlin.
00:00:36: Welcome, Izzy.
00:00:37: As always, we start with quotes before introducing our guest.
00:00:40: Of course.
00:00:41: And maybe I can already hand over to our guest for the first quote.
00:00:50: Yeah, happy to give the first quote.
00:00:52: Welcome everybody here.
00:00:54: The first quote is, we are optimized for local short-term decision making, yet we are having collectively as human race.
00:01:03: global and long-term impact on our planet.
00:01:06: And so the microscopes that we are going to discuss today help us understand more global long-term patterns, strengths, outliers, and also help us to invent together desirable futures for everyone.
00:01:21: Another one that you found easy that I really loved was If you torture the data long enough, it will confess to anything by Ronald Koos.
00:01:33: I love it.
00:01:34: Yeah, I love it too.
00:01:35: I think, and we'll get into how to not torture the data too much in order to get actually meaningful information out of it instead of torturing it into telling us a lie.
00:01:46: I will finish also with a quote from Otto Schopenhauer, which says, it is only in the microscope that our life looks so big.
00:01:54: And I think this is very... nice because, as Katie introduced, we'll talk about macroscopes that help us see the bigger picture, see complexity.
00:02:04: And this is an analog to a microscope that actually allows us to see our life in full scale.
00:02:11: And I think we're going to have an interesting discussion about that today.
00:02:15: And maybe we can introduce our guest, Katie Berner, and I will.
00:02:19: introduced the many accolades that you have by reading them from my notes.
00:02:24: So Professor of Engineering and Information Science at Indiana University and the Founding Director of Cyber Infrastructure for Network Science Center.
00:02:34: So this is on the science side.
00:02:35: Then I have another science part, which is the human reference atlas.
00:02:39: And maybe we can touch upon that very, very briefly when we go through the microscopes.
00:02:46: But you are also the curator or one of the curators behind places and spaces, which is I think started as a data visualization or science visualization exhibit and also morphed into macroscopes.
00:03:00: So I think we can talk about that.
00:03:02: And also host of envisioning intelligence, the twenty four hours event where I could also partake.
00:03:09: Very interesting talking about.
00:03:12: many different levels of intelligence from the artificial to the human to the microbial and anything.
00:03:20: Yeah.
00:03:20: So I think you're wearing many hats.
00:03:23: And I think there is also an artistic quality to what you're doing, right?
00:03:28: So you don't see science, I would say only as a discipline of numbers, maybe, but also as a discipline of pictures of images of something tangible that needs to be explored.
00:03:41: Absolutely.
00:03:42: Happy to be here.
00:03:43: Thanks for having
00:03:43: me.
00:03:45: So yeah, maps, microscopes, and the third iteration of the exhibit is on envisioning intelligence.
00:03:51: Yeah, thanks for joining us.
00:03:52: Invisioning
00:03:53: intelligence, I like that.
00:03:55: So maybe maybe we try and distinction between introducing the show.
00:04:01: I said, we're not talking about telescopes, not microscopes.
00:04:06: How Do I look at macroscopes?
00:04:09: for those as myself who don't have a clue what a microscope might be?
00:04:14: It's not a thing.
00:04:15: I understand.
00:04:17: It's beyond the thingy stuff, is it?
00:04:21: Oftentimes it is a tool, but it's also a process.
00:04:25: So you take data, which is oftentimes massive and complex and complicated and multi-scale, and you try to make sense of it for yourself, for those surrounding you, for your work, for those which depend on you, if you're a doctor and you need to make decisions about patients.
00:04:45: And you try to not only make sense for yourself, but you also try to explain to others what the data means, what kind of trends, outliers, patterns are in the data.
00:04:54: And ultimately, how you can now get to meaningful actions, be it to stratified patients, be it to cure patients, be it to do precision helps instead of precision medicine.
00:05:07: Ideally, you don't even get sick.
00:05:10: So I think it's a really interesting way to Look at the data you're surrounded with and make better use of.
00:05:17: Yeah.
00:05:18: Is it kind of a starting point that the assumption is we are kind of overwhelmed by data right now because AI, for instance, produces such massive amounts of data that we are not capable as human beings to make sense of it anymore.
00:05:35: And then you come to the aid.
00:05:38: Yeah.
00:05:39: So in former times, it was those which had the resources that could make sense of data.
00:05:45: So investigative journalism, for instance, or those which had access to big data streams and the compute power to do that.
00:05:52: But more and more, you can give these tools to anyone and everybody can then make sense, make their own sense of data, try to understand if it's true data or if it's falsified data.
00:06:05: if it's data that you need to act on or somebody else needs to act on.
00:06:08: And also what actions are best for all of us, not just for you personally, but ideally for many, because if just you benefit in the long run, that won't work.
00:06:19: True.
00:06:20: How would you see so, because you mentioned AI, right?
00:06:23: AI is a tool to make sense of data, but you have like a black box in between.
00:06:29: So you have an input and then you get some kind of prediction as an output, but you have no idea about the middle layer.
00:06:35: How would you compare that to a macroscope?
00:06:38: That's a great question.
00:06:40: So first of all, you can visualize the training data that's used to train an AI system.
00:06:47: You can visualize the architecture itself that this system consists of.
00:06:52: You can visualize what that large language model or large vision model can generate or not.
00:06:58: You can also help us all understand what happens if you add new training data.
00:07:02: so all of this can be visualized and will be visualized in the third decade of the exhibit.
00:07:07: But the microscope tool typically lets the human decide what are the stakeholders, what do they really need to understand, what is most useful for their decision making and then you would go about getting your hands on the best highest-quality data, analyzing or modeling that data, visualizing it, trying to convert it back to something that the decision-maker understands, because oftentimes it's like a mass problem.
00:07:32: You have to take a real-world example, convert it into something that you can solve as a mathematical equation, and then convert that result back to the decision-making process.
00:07:42: And so, same here.
00:07:43: You would go through this entire process, and oftentimes you just realize, oh my God, I missed data, or I really need to to a different analysis.
00:07:51: I don't want to just see trends.
00:07:52: I want to see a network effect.
00:07:54: I want to see geospatial diffusion patterns.
00:07:57: And so you go this feedback cycle one more time.
00:08:00: And then the decision making, hopefully, is much more informed and not just local and short term, but more global and long term.
00:08:10: So you learn something about your data while you work on the microscope instead of just putting something into the AI and hoping for a correct answer, which I think we, I mean, for a long time, I think people didn't really have access to work with AI, right?
00:08:25: So it was like more of a niche, but nowadays that everyone can work with JetGPT co-pilot and all the other related tools.
00:08:32: I think we're really firsthand.
00:08:35: Everyone is experiencing this frustration of having a big data driven output that is generated by AI, where you simply don't know the uncertainty of it.
00:08:47: So you have the disclaimer, I can make mistakes, I'm a large language model, I will not know the truth.
00:08:53: But
00:08:53: you don't really know how true the answer might be.
00:08:56: And I think having this working with the data, so knowing what is all the input that went into it and how much can I trust this data, because I think this is really important, right, to know.
00:09:07: is this data that Yeah, has been curated that has been replicated or is that just some data set that was even produced by an AI for example?
00:09:17: So I think that is it's a really nice layer of what a microscope might be able to do.
00:09:23: Yeah, totally agree.
00:09:24: and you can also use large language models to create a data frame to identify high quality data sources and to then analyze the data.
00:09:34: And many of the large language models also help you create geospatial maps or trend lines or even network layouts.
00:09:42: So I think as we get more and more functionality to come into existence, you can go through that entire cycle of identifying a data frame, the best data possible, doing all kinds of geospatial temporal topical analysis to creating data visualizations and helping us all understand if those are correct or incorrect.
00:10:04: To the point of impacting daily decision making, all of this can be supported by AI.
00:10:09: Absolutely.
00:10:10: That's
00:10:11: very fascinating.
00:10:12: Amazing even.
00:10:13: But at the same time, I asked myself a question in the process of curation.
00:10:19: How iterative is this?
00:10:21: How much people power is invested in the process?
00:10:25: What is your experience in that sense?
00:10:28: Ideally, there is always a human in the loop, at least in the beginning, until we know that it's a hundred percent correct.
00:10:34: We're talking about one human.
00:10:36: Well, and oftentimes it's multiple because the person who has the database skills is oftentimes not the person who has the people's skill or the person who has the design skills.
00:10:46: So oftentimes there's an entire team effort needed.
00:10:49: But ultimately, it is important for us to understand.
00:10:53: in how far there is a hallucination going on, as you mentioned, or if there is a truth in the making.
00:11:01: And oftentimes you can triangulate, right?
00:11:03: You can ask an expert, you can ask a literature, you can ask a lot language model.
00:11:07: And if all three of them point in the same direction, then most likely that's gonna be true.
00:11:13: But sometimes all three are wrong.
00:11:24: I think one interesting part... about the macroscopes also that we deal a lot with storytelling.
00:11:32: So I think this is something that we discussed a lot about over Christmas, right?
00:11:37: In the family space about the future of school, the future of learning.
00:11:42: I mean, in the age of AI, what are the skills that young people need?
00:11:47: And we came to the conclusion that storytelling is really tremendously important because this is the way we make sense of the world.
00:11:54: And this is a skill that we will always be able to use independent of which jobs or which task AI will take from us.
00:12:03: So, and I think this is a really interesting aspect about a well done microscope that this includes this element of storytelling through the data.
00:12:13: And I wonder if you if you could comment maybe on an example over how you feel about this layer of the picture.
00:12:21: Yeah, I think stories are powerful.
00:12:24: We think through stories, we remember through stories, we like to communicate through stories.
00:12:29: There's an entire movie that an artist, sound artist and myself created, which is called Human Access.
00:12:36: It's online, twelve minutes, communication through the ages, how we use stories to communicate to each other and to invent desirable futures.
00:12:44: There's kind of three futures in the movie depicted.
00:12:48: But in these data visualizations, several of the microscopes, so the second iteration of the exhibit, are actually scrolling stories.
00:12:56: So they tell stories.
00:12:58: as you go down a website.
00:13:00: You start from the top, which has an interesting title.
00:13:03: And maybe some X rolling by, the scrolling story for the X shape.
00:13:07: And then you go down and you get more and more information, more and more examples, more and more interactive data visualizations.
00:13:14: And the New York Times and others have also used extensively scrolling stories to explain the data.
00:13:21: explain the purpose of the data analysis itself and then go down to help people understand how this all works and what the take-home message is.
00:13:31: And as we go into AI, there are a number of really nice scrolling stories also that explain how these AI algorithms work, which I think we need to understand better so that we can work with these intelligences better and understand when to trust them and when to not trust them.
00:13:47: as I get a teenager kind of.
00:13:49: Such an important point with the making AI explainable through this kind of the example that you just mentioned.
00:13:56: I find it fascinating that the X shape example, I think it wasn't science magazine, right?
00:14:00: We looked at it yesterday.
00:14:01: And to
00:14:02: be honest, I'm not that invested in X shapes.
00:14:05: I don't care that much.
00:14:06: Yeah,
00:14:07: but I was really invested in the data visualization in the way this story was told, because I have never found myself before thinking.
00:14:16: what might be the egg shape of this bird and why are they so different and why are they not so
00:14:22: round.
00:14:22: Yeah,
00:14:22: I forgot.
00:14:23: The tiny, tiny bird round shape.
00:14:26: Not the ones which are high up on the cliffs.
00:14:28: Because if you have a nest there and one egg falls out, it has to roll in a circle.
00:14:34: Otherwise it just rolls
00:14:35: around the cliff.
00:14:37: So if you have a roundish egg or a chicken kind of egg, it would just roll on.
00:14:44: So I think there's a reason why these X shapes exist and they just make a very beautiful story.
00:14:52: And I think this is cool because it also shows us that if the data is presented well, you immediately care about it.
00:15:00: And this is great, right?
00:15:01: Because it speaks to the soul.
00:15:02: in a way, it speaks to the human intuition of wanting a story, wanting maybe also causality, as you just say.
00:15:10: I mean, there's a reason behind the X shape, and suddenly I care about the X shape, which I really didn't do before.
00:15:15: So I think this is a nice way of seeing
00:15:19: how...
00:15:19: this huge data that is very intangible for us, right?
00:15:23: It's not something that we can comprehend that we can consume.
00:15:27: And I think we noticed this in the pandemic times a lot that people were really struggling to deal with these big data sets and with just making sense of what big data might tell us.
00:15:40: And I think having it presented in an easily explorable way and also where you as a user can do something with the data.
00:15:50: I think that makes it really, for me, much more interesting.
00:15:54: And I was wondering, is that something?
00:15:56: because, I mean, you, how many years have you done the places in spaces, fifteen or something?
00:16:00: Well, it's twenty-one years.
00:16:02: We just have the first iteration of the third decade.
00:16:05: Cool.
00:16:06: But then legal
00:16:07: age.
00:16:07: Do
00:16:08: you see a difference in how... people react to the data if they can explore it as compared to just viewing
00:16:18: something?
00:16:18: Absolutely.
00:16:19: So the first decade was really on maps of science, maps of our collective scholarly knowledge.
00:16:23: And those maps of science had not been seen outside of national labs and maybe some other institutions which have again the riches and the foresight to do that.
00:16:33: In the meantime, these maps of science are in many, many places, they are commonplace.
00:16:37: So the exhibit really just opens the door and said, here, this can be done.
00:16:41: And there is no argument that you could possibly have if you see a hundred maps of science that this can't be done any longer.
00:16:48: And so it also introduced new areas of science.
00:16:51: So some people just printed the map with nanoscience or neuroscience, showed them to their dean and said, here, I exist.
00:16:57: I exist on this map.
00:16:59: We should fund this.
00:17:00: We should hire.
00:17:00: We should have students in that area.
00:17:03: And then also because of the pandemic, because of COVID, many people learned about logarithmic scales.
00:17:10: They learned about flattening the curve.
00:17:12: They learned about pandemics and delays in the system, right?
00:17:15: So you have an outbreak, then the hospitals flow over, then their death.
00:17:19: So I think these delays in the system, they are super important to understand, feedback cycles super important to understand.
00:17:25: And so the data visualization literacy really has improved enormously because of COVID.
00:17:31: And that's a sad reason, but it's a good news item because we need to understand our world better.
00:17:37: And it's also really important to have remarkable stories like the X story, because then you make remarks to your family members over Christmas and the good news spreads.
00:17:48: Otherwise, it's mostly the bad news which spreads, and especially also misinformation, because it can be fabricated in a way that it pushes all the human buttons.
00:17:58: and send that spreads.
00:18:00: And what we actually do want is that truth and wisdom and good advice spreads.
00:18:05: And that's really a challenge still.
00:18:17: Talking about the storytelling aspect, as you might imagine as a lifelong journalist, I would have a zillion questions about that.
00:18:25: But I want to reduce it to maybe two.
00:18:30: One of them in our age and time where it appears to me that we need translation efforts from science to politics, from science to civil societies, to culture, whatever it is.
00:18:46: Do you think you can essentially help in this process where like geopolitics seem to seem to be very complicated right now and even increasingly so that we need especially this translation process.
00:19:03: We have worked extensively with science policymakers but also technology policymakers and others like local regional development offices.
00:19:13: They all are interested to understand what is the next big science or technology or innovation because they want their regions, their science their foundation, their stiftung to succeed.
00:19:26: And so they really have embraced over the last years this ability to take rich data.
00:19:31: And in the US, most of the publication data, patent data, funding data is all available online.
00:19:37: So you can mine it also using again, lot language models, lot vision models, you can go and even harvest all the figures.
00:19:43: And then these policy makers can empower themselves by doing more data-driven decision-making, really using that rich data, that high-quality rich data to help us all understand what teams or what efforts, what areas of science to fund with what kind of mechanisms.
00:20:00: It makes a difference if you fund a workshop or an all-one investigator-initiated project or a big common fund project.
00:20:08: It's not always clear which one is the best.
00:20:10: And in some cases, you really need big science.
00:20:12: If you want to map the human body, you need big science or the human genome, big science effort.
00:20:18: But if you have something very specific, which one good team can solve, then maybe they are one mechanism, a superfect way of finding that.
00:20:27: Yeah.
00:20:28: Easy if you allow the second question.
00:20:30: Of course.
00:20:30: The story telling corner of the world.
00:20:34: I'm listening to you and.
00:20:36: I got increasingly interested in your personal path to macroscopes because, you know, as an academic at the university, I imagine it's a comfy place to stick with the data, stick with algorithms, stick with the math and not leaning out of the window.
00:20:56: Of course, in a sense, becoming a little vulnerable as well.
00:21:01: So why did you do it?
00:21:03: meaning the exhibit or the twenty-four hour event or working as artists.
00:21:09: Everything, basically, if you have a short version of that.
00:21:13: Short might be tricky, but twenty-one years ago we had a wonderful breakfast and we thought, well, few people really know about these data visualizations.
00:21:23: And we thought, well, by framing it and bringing it to many public places, libraries, museums.
00:21:28: other university conferences, that would make a difference.
00:21:32: And it did make a huge difference.
00:21:33: And we learned about new possibilities and we got in contact with some of the coolest people out there making these maps and some of the really neat new data sets and algorithms and tools.
00:21:45: It's a fun life.
00:21:46: I highly recommend doing that.
00:21:48: But it's also named science communication.
00:21:50: And I think there is a true value in the second translation process, right?
00:21:54: So what we are doing here is to try to take somebody out.
00:21:57: out of the university setting and bring it to people that might already be interested in data visualization, medical precision medicine and precision health.
00:22:06: But ultimately there needs to be one more step, right?
00:22:09: So that my mom, my children, all of us can ultimately understand it and empower themselves with these new tools and toys.
00:22:17: And it's fun.
00:22:18: It's not just a chore that you need to do.
00:22:20: It's fun to look at your own life, at your own data and new ways with new eyes and to well make better decisions for yourself, for your family, for your environment, for your neighborhood, for your country ultimately and for your planet, for your planet you have.
00:22:36: And so I think there is a beautiful feedback cycle in there between going out, getting to see the real needs and then bringing those back, building tools as you asked about my career trajectory.
00:22:50: So I studied electrical engineering in Leipzig and That really made me into somebody who wants to serve the needs of others, build bridges, build airplanes that don't fall off the sky and empower them to do the things that they would really like to accomplish in life.
00:23:06: And tools have always made us more powerful.
00:23:09: Yes.
00:23:10: Now we are building data visualization tools or frameworks, ontologies, common coordinate frameworks for the human body, because that ultimately will help us understand life.
00:23:21: understand health, understand ourselves better.
00:23:24: And it will always be tools that help us understand the data better.
00:23:28: We can't just take our bare hands and our bare
00:23:34: brains.
00:23:38: To do that.
00:23:39: We can't if we're not made for that.
00:23:41: Similarly, we can't get a nail in the wall without a hammer.
00:23:55: Where would you say Do you still see challenges with these types of data visualization?
00:24:03: We talked about data uncertainty, and I think especially in the biomedical space, you have a lot of just variability that comes through the variability of the individual, of the individual cell, of the timing.
00:24:19: There is so much.
00:24:22: real noise in the data, biological noise, that's not an artifact, it's simply there.
00:24:27: How would you say,
00:24:28: I mean,
00:24:29: can you do justice to visualizing the full complexity, maybe in a biomedical example, but also in others, I think also other areas are complex, I just care most about biomedicine.
00:24:42: And at the same time, make it simple enough to be explorable, to be interesting, to not overwhelm.
00:24:50: the user, the explorer.
00:24:52: There's a huge difference between mapping science where you have papers and patterns and grants and job advertisements etc.
00:24:58: and that's basically set in stone.
00:25:00: and maybe there are retractions or corrections but ultimately you have a good record.
00:25:05: For the biomedical domain things change.
00:25:09: It depends on what you eat in the morning or how well you slept, what kind of blood values you're going to get in your frozen blood build on the next morning and also it will depend on your cycles be it menstrual or daily or whatever of what you will see in that data and our understanding of the full complexity across all temporal scales from less than a second to lifelong and all spatial scales from meter size human body to nanometer protein sizes.
00:25:45: This complexity, I think, is very little understood.
00:25:48: And oftentimes, in order to understand it, you would have to have interventions, like you would have to ask somebody to lay down, stand up.
00:25:55: Many things have to go right, and not wrong, to accomplish that feast from laying down to standing up, or just going right upstairs, or having some glucose intervention, a glass of orange juice or carrot juice.
00:26:10: All of these interventions then will help you understand the dynamic reaction of your body to these external shocks.
00:26:21: That kind of data will be needed to really help us understand how to model the human body and ultimately also how to map it and understand it and keep it healthy and to cure disease when it comes into existence.
00:26:33: In the Human Reference Atlas but also in the Human Physiom Project that we just started, we are trying to get our hands on that kind of data and that in many cases involves collaborations with the best teams out there which really do these interventional studies and ultimately taking that data across temporal and spatial scales and across all the human physiological systems to help us understand how this all interrelates.
00:26:57: And in many cases, it's feedback cycles that keep us in a green steady state so that we don't have too high blood sugars or have too low or too high calcium levels and so on.
00:27:10: And so I think we are getting closer to that, especially also with AI, which can help enormously to help us decide for life.
00:27:17: But we need these interventional data sets.
00:27:20: And only then we can identify what keeps us in the green range.
00:27:26: There's a green camera.
00:27:29: So something that we thought about when we discussed beforehand for this episode is, do you know if macroscopes are being used in a business setting?
00:27:38: Because I mean, so the way that I have read it.
00:27:43: It's a tool for decision-making based on data, right?
00:27:46: And there are a lot of people, I mean, just thinking about biotech and pharma who want to make decisions based on data.
00:27:52: For example, is it a smart idea to invest in this compound, into this disease area?
00:27:57: Should I continue with a clinical trial based on the adverse event profile?
00:28:01: All these kinds of questions.
00:28:05: have come across macroscopes in the context of scientific literature.
00:28:09: So I think nature and science are doing it a lot also for visualizing what they have in stock, which I find very cool.
00:28:16: I mean, they have a lot of nice ones.
00:28:20: But in a business context, I come across dashboards much more often than I come across macroscopes.
00:28:27: Do you know if people are using it in a business context?
00:28:31: So in the Atlas of Knowledge, which is the second in the trilogy, you have quite a few examples of maps for science policy makers, maps for economic decision makers, that would be the ones that you would like to look at, maps for scholars, and then also for general audiences.
00:28:48: And so for the economic decision makers, you absolutely do patent analysis like, what is my intellectual property landscape?
00:28:57: what are the other companies that I care about and try to compete with on the global market doing.
00:29:04: You could also use job market data, like what are they hiring, what should maybe I also consider hiring, what kind of talents.
00:29:11: You also want to understand delays in the system, like how long does technology adoption.
00:29:17: need.
00:29:17: You might also like to see who's close by where you can then make collaborations or hire the best students.
00:29:24: I think it's ultimately about talent trajectories.
00:29:27: You want the best people to work in your environment or the best AI to work in your environment now soon.
00:29:33: And also ultimately supply chain optimization.
00:29:36: So that's more global maps, geospatial maps.
00:29:39: But then also more and more of these scientific maps are just disciplines of science which have to come together to support, for instance, the creation of the videotape recorder.
00:29:49: There were fifty distinct technological events which were needed to create that device, which very few of us now use.
00:29:57: Almost extinct.
00:29:58: Exactly.
00:29:59: But then imagine how many technological events are needed to create an airplane.
00:30:05: It's amazing and so understanding this better bringing that together so that you can be the first on the market.
00:30:10: I think it's really important.
00:30:12: It's academia.
00:30:14: ish still a little bit as it sounds to me.
00:30:17: It's not as close to real business real world business as I would imagine this to be necessary because.
00:30:27: As you laid out, I think in each and every decision, this should be a part of it.
00:30:32: Not only in science or politics or science-oriented businesses that are like startups.
00:30:41: This should be incentivized in my thinking.
00:30:45: What I hear now, it's like an offering, but it's take it or leave it.
00:30:52: Yeah, in a sense.
00:30:53: I
00:30:53: would think there may be three different types of usages.
00:30:57: One is you have a dashboard, like the COVID dashboard, which many of us went to regularly just to see what's going on.
00:31:04: Obsessively,
00:31:05: true.
00:31:06: Obsessively.
00:31:09: Then there is the kind of decision making room, like many banks have going way back in time, stock market.
00:31:16: charts and they regularly go there to remind themselves what has happened a hundred years back in time.
00:31:23: and then there are the very special exercises where maybe once every two months or maybe once every half a year you do this more global analysis and that's more complicated and maybe not everybody gets to see it.
00:31:37: but that then also looks at geospatial spreading at network effects at feedback cycles at causal loops, et cetera.
00:31:45: And I think that maybe is not yet everywhere because it's so complicated to read.
00:31:50: these network visualizations and dashboards are just so much easier to read.
00:31:55: But I think there are these three stages and many of the big companies, they do all three.
00:32:00: They have the daily decision support.
00:32:02: They have the go to get together and look at trends and outliers, et cetera.
00:32:07: And then the more complicated deep dives into the data.
00:32:24: And I think the last one, so when we talk about AI, we have talked a lot about encoding human intuition.
00:32:30: So how to, I mean, this intangible aspect of decision making, you have seen the data, but somehow your gut tells you why you want to go into area A instead of area B. And I can imagine that for the microscope, for this complex data deep dive that takes a little bit this way we process data as a story that this can actually be a really good decision making tool because it brings the data closer to maybe the way that our cognition.
00:33:04: really works and I think this could be pretty cool.
00:33:07: And
00:33:07: might convince the CEO not to decide.
00:33:11: Not to
00:33:11: do like.
00:33:13: Exactly.
00:33:15: And to be honest, the computer scientists and engineers, which oftentimes do the data visualization, they might not be the best storytellers, right?
00:33:22: So oftentimes there needs to be a marketing expert or somebody in the mix as part of the integral team.
00:33:29: That then can tell that story to the person, to the sea level, which then can make that decision.
00:33:35: Ultimately, however, if you can empower those decision makers to make their own analysis even better, and my best example is cancer boards.
00:33:44: So you have a patient, a cancer is in progress, and now you have different people sitting around a round table.
00:33:52: looking at all of the data with different areas of expertise and trying to make the best possible decision.
00:33:57: So oftentimes it's not just one person, it's really a team of experts together with CAI also and the data visualization and a lot of tabular data also.
00:34:06: It's not just the visualization.
00:34:08: Trying to understand what is the best possible treatment because there might just be one chance of life.
00:34:16: That then has to be communicated to the patient as a story so that the own ambition and hope also can help to get healthy and back to normal again.
00:34:25: That's a great example, actually, of how many decision-making processes should work in the best possible way.
00:34:31: Who do you think should be involved?
00:34:34: I mean, we talked a little bit, you mentioned some of the disciplines already, but I think, I mean, I have the feeling that if it's well done, like these examples from science or from nature.
00:34:44: There are very artistically skilled people involved.
00:34:48: There are people who are skilled in storytelling and there are people skilled in data, right, who are in the mix.
00:34:54: Yeah, that's a good start.
00:34:56: So every spring, I teach a course to data science, computer science, and then students from all kinds of different disciplines at Indian University.
00:35:05: And for those projects, we have client projects.
00:35:08: Those are projects which have a great data set, a great question, but no money.
00:35:12: And I can't use my professional staff because they want to have a salary at the end of the month.
00:35:16: So I just can't use them.
00:35:18: But we then incentivize these potential clients to just submit this as a course project.
00:35:25: And then our students and teams of four to five work together on these projects.
00:35:29: In those four to five teams, they need to identify somebody who really understands the data, really, really understands it, and works with the official data provider to get even more data.
00:35:42: and information.
00:35:43: The next person has to really have a deep understanding of the data analysis and visualization algorithms that exist.
00:35:49: The next person has to have some kind of aesthetic and design, image composition, color selection, all the graphic variables and spatial variables that you want to use.
00:35:59: And then the next person really has to be the contact to the client.
00:36:03: They need to really talk with them, keep them in the loop, understand what they really need in their life, not just faster horses like Henry Ford's example of automobiles.
00:36:13: But what they really would most benefit from, and that of course requires that they understand what the possibilities are, right?
00:36:20: And then the last one has to document.
00:36:23: And of course, ideally all of them document, but there might be one that just takes the lead and inspires and encourages the others to document, because otherwise it won't be going well.
00:36:35: And so if you have this kind of team, then I think a lot of magic can happen.
00:36:39: And it's really fun for these students to then have these cool portfolio items where they really made a difference in somebody's life or in some work process.
00:36:47: And so ultimately, yes, there is a story at the end and a written up paper.
00:36:52: But it's also important that there are these different elements that come together.
00:36:57: And there are maybe two or three of our macroscope and map makers where that all is in one person.
00:37:05: But these are Very much exceptions.
00:37:08: Typically you really need a team to do it, good data visualization.
00:37:12: So you would say it's worth it to invest for companies into this effort to get a good team or... get help.
00:37:21: I think so.
00:37:22: And many companies have internal data science courses and data visualization courses.
00:37:27: They want to flexi data muscle, whatever they call that, and ultimately also empower themselves to make better decisions.
00:37:36: because if they don't make those decisions, then other companies are going to win out.
00:37:42: Easy.
00:37:42: Extremely fascinating.
00:37:44: This episode, what's your take out in one sentence?
00:37:48: Oh, I hate one sentence.
00:37:49: Okay.
00:37:51: We come in again.
00:37:52: My one sentence.
00:37:53: You need stories to understand data.
00:37:56: I think this is my one sentence.
00:37:58: My takeaway.
00:37:59: We leave it with that.
00:38:00: And this was the Bio Revolution podcast.
00:38:03: Please subscribe to us at Spotify.
00:38:05: Leave a comment if you want to.
00:38:07: And further information, as always, you'll find at ScienceTales.com and the tales.
00:38:14: I mean, this is a running gag.
00:38:16: I repeat it once more, come on.
00:38:17: T-A-L-E-S.
00:38:20: It's not the tales of the dog, but the tales of the fairy tales.
00:38:23: Kelly, it's been fascinating to have you here in the show.
00:38:27: Very many thanks.
00:38:28: Great studio.
00:38:31: And maybe you could share the atlas and the places and spaces that are upcoming.
00:38:36: We will link them in the show notes, but for our listeners and viewers to find.
00:38:41: Yeah,
00:38:41: so the Atlas of Microscopes just came out.
00:38:43: We're still on a book tour.
00:38:45: In fact, I'm going to count this as part of the book tour.
00:38:48: So we went to many different countries and places, and Berlin is just... cool place to add.
00:38:53: We do have the exhibit online at simapps.org.
00:38:56: And so you can find a hundred maps of science technology and also for kids.
00:39:01: There are science maps for kids in there.
00:39:03: And of course, forty microscopes.
00:39:05: These are interactive data visualizations that you can explore.
00:39:09: And very soon you will get to see envisioning intelligences.
00:39:13: This is the new exhibit that looks at how we can make better sense of other types of intelligence.
00:39:21: and how we can interact with it and use it and make the best of it.
00:39:28: Fascinating.
00:39:29: Now we need another forty minutes to explore the intelligence.
00:39:32: We
00:39:33: definitely do if that's sufficient.
00:39:36: Many thanks again for listening and watching us see you around at the Biorevolution Podcast.
