Seminar with Clancy Wilmott "Data practice as theoretical inquiry..."

I'm Kitskar. I'm the executive director for BIDS, the Berkeley Institute for Data Science. We are a space um that is invested in making it easier to do meaningful interdicciplinary collaborative work for the betterment of society. We've had some really interesting pre-con conversations about how hard it can be sometimes to get the technology working in service of the interesting research question. And um so I'm really delighted that everyone is here. Thank you to everyone who's come in person. Thank you to everyone who's joining us online. Thank you particularly to Lily who um coordinates our events and sets everything up. And thank you so much to Clancy for coming along and presenting to us. So we will have um around about 25 minutes 30 minutes of presentation and then lots of time at the end for questions. So please ask your ask your questions then um I will run around with the microphone for you. If you are online you can ask your question by typing it in the chat. So please also participate and ask those questions um if you're joining remotely. Thank you again everyone for coming along. Um I am delighted to introduce Clancy Wilmott. uh she's an assistant professor in critical cgraphy, geo visualization and design in the Berkeley center for new media and the department of geography. She comes to UC Berkeley from my hometown which is uh Manchester where uh Clancy completed her PhD in human geography with a multistight multi-sight study on the interaction between mobile phone maps, ctographic discourse and postc colonial landscapes. Um, I'm going to just say one more thing from her bio. I'm not going to read everything, but it's so sort of wonderful to have her here because we are really investing in celebrating people who are coming from different disciplines and different backgrounds. And um Clancy herself sort of integrates these different backgrounds because she holds undergraduate degrees in communications, media arts and production and international studies in Italian as well as a post-graduate degree in cultural studies from the University of Technology in Sydney. So I am absolutely delighted that you're here. Really fascinated to hear more about data practice as theoretical inquiry. And please everyone would you join me in welcoming Clansley. Thank you. Hi. Should I say good day? Um, thanks for having me. I'm really keen to talk um today and I suppose propose some ideas that I've been thinking through and working on over the past few years. Um, my apologies. This talk may be a little bit of a best of hits uh kind of discussion um as I go through some of the projects I've been working on, but uh I feel it's necessary in order to invoke um examples of what I'm understanding as data practice as a mode of theoretical inquiry. And I mean theoretical here in the humanistic sense um in the sense that we're sort of thinking about scale. We're thinking about interreations between objects and processes. We're thinking about politics um both in terms of experience and knowledge. Uh and so much of this work has come from a fairly unenviable position that I found myself in throughout my career uh which is being stuck in the middle of a disciplinary war. Um and so this is happening in geography. Uh during my PhD I was part of a research cluster called mapping uh culture/GI science which combined cultural theorists uh mostly forodians with uh physical scientists who mostly focused on pollen sampling and environmental processes. Um and this group pretty much encapsulated the two ends of the debates that we find in geography around ctography and geographic information science where on one hand we have humanist critiques that GIS and ctography is inherited colonial discourses and are imated with capitalist economies and are built from the military-industrial complex and can be used as technologies of control. So this broad uh based both Marxist and post structuralist critique. Um but on the other hand what we know from those who work on the ground those who are environmental scientists and social scientists that mapping GIS and ctography are also really invaluable to landback movements to grassroots environmental monitoring and management to forms of spatial storytelling forms of counter mapping and crisis management. So there's actually this whole range of really important and useful ways in which data itself spatial data uh can be used. um to help people uh improve their lives or deal with modes of crisis. And so stuck in the middle of this is someone who both uh engages with and reads the critical theory around ctography and power but also routinely uses data science and spatial data science to help communities has been a very uncomfortable experience um as I try and navigate um two sides of a of a debate um where there's not a lot of room for uh consiliation. So I have approached this in my own teaching research by first asking not necessarily what are the politics of any particular project accepting that all data projects have a politics but really asking where are the politics where in the data process do politics lie and so for me um that may be in the fact that a grassroots mapping organization is using a piece of commercial software like ArcGIS to gather pretty basic data that will be given to a government in order to advocate for better health outcomes due to the location of a power plant nearby. And the wear of the politics in that case may not be in the forms of counter data or open-source mapping or necessarily the creative outputs that it comes from, but the fact that people are creating maps for themselves. The where is located in the people who are mapping. In the same way, sometimes the process itself is enough. the process of describing. So we see this a lot in participatory mapping with indigenous communities where the output is not necessarily the point so much as the point of claiming land and describing it and creating ontologies that actually make an argument about how the world is for a certain group of people. Similarly, sometimes the politics are in the output are in the in the process of using open- source maps of open source data um of making that available um and actually thinking about countering the increasing corporate control over both our data itself but also the way in which we manipulate or analyze data as well. Um and then sometimes it's in the output. Sometimes it's a question of who gets that data. Sometimes the data does not end up in the hands of um a government but instead ends up in a server somewhere um that a community can use. Um and I know a number of colleagues who have undertaken data work that they've never actually published or been able to publish because the point of that the wear of the politics is in the community itself um not in its academic impacts. So throughout all of this um I've been thinking about um this question as well of of excess in a time of big data. Um and so I was accused recently of being a small data scientist. Um which I feel like is a harsh accusation. Um I very rarely use data in the millions. Um I very rarely use sort of large scale. But I've been thinking as well about this question of excess um and how we seem to have so much data about the world that sits in excess of often what we really need to know. Um and so I've been thinking through this question of can this this excess this space of data that sits actually be used to help us challenge our axioms not just within data science but actually within the arts, humanities and social sciences. um specifically from a geographical perspective arguing if the different epistemologies and ontologies that have been developed out of specific cultures do they serve those spaces better. So the excess in this case is both the excess of data but also the excess of the world that does not get captured by data itself feelings stories memories etc. Um secondly, what doesn't fit on the map or is not in the system also offers unique insights into the limitations and rigidity of those sames maps and systems. So that which data cannot capture also modes or creates a mode of critical and theoretical inquiry. So not just how data is constructed but also um what exists beyond data and how that world works. And then finally, I think for me the most important question that I ask is do we have to accept the convention, the axiom, the datim, the base map, the metric, the lack of interoperability in the object. So do we have to accept our current data present? Um, and I talk about this specifically with regard to geographic information systems and spatial data, which during the 1990s seemed to have fairly unquestionably just inherited and adopted a whole range of geopolitical systems um without asking whether or not they were still fit for purpose. And that comes from scalar systems, so um states, counties, etc. through to descriptive systems as well. categorizations environmental versus social as if they are separate kind of um domains. So what does this look like in practice? I mean these are all great ideas. Um but I wanted to again like I said focus on some examples so that you can so it's possible to try and imagine how and where those politics might emerge and the kinds of questions I causes to ask. Um, and so this mode of thinking in its first instance developed out of a project that I was working on with the Wood Street Commons, which was a group of unconventionally housed people who were living under the I880 in Oakland, uh, in West Oakland. Um, and they approached my studio with a pretty simple question, which is they just wanted a map of who owns the land under the I880. And I was like, "Okay, super easy. Fine." And so, not having worked in the US data sphere before, I was like, "Yeah, give me a day or two. I'll get that to you." Yeah. Funny. Um, so several weeks later down at So, basically the way in which this space works um is that it's deeply fragmented. Um, and so there are about 46 to 48 parcels of land under the I880 which have been mostly fractured by different agreements between the state of California, the city of Oakland, and different railroad companies um that have merged or not merged or fallen apart. Like it's a complete mess. The issue is that if you want to find out who owns the plot of land, it's actually not available on the internet. you have to call up the um Alama County Assessor's Office and ask for each parcel individually, but you're not really allowed to ask for more than eight or 10. So, if you're living under somewhere, like not over the phone anyway. They get annoyed. They got very annoyed when I'm like, "Okay, so I've got 47 parcels." They're like, "Uh, no." So, like, "Okay, I'll come in and look it up on the computer." So, we went in and looked it up on the computer. And to do that, we had to go through a metal detector and sit on a computer that looked like it was from 2005 um and look up each of these parcels individually. And I was getting a little bit exhausted by parcel number 24 as we're sort of hand writing this information because it costs to print out every single one. And I did something that I should have done in the first instance, which is contact our GIS data librarian and say, "What am I doing wrong?" And she said, "Oh, we pay for that data. Here it is and she sent me a zip file with the names and addresses of every single parcel owner in uh Alama County and then in about an hour I put it on a map and sent it down to Wood Street. the fact that I was able to access this data when I didn't need it compared to the fact that for your average person who is unhoused um to be able to or to have to go into an office um and sit there while a security guard who is pretty burly watches you sort of raised a whole range of questions about the different forms of politics that we have over uh urban margins. Um and we ended up publishing a paper sort of pushing it back against this idea that property logics invisibleize um people who are unhoused only and instead argued that actually property logics also invisibleize those who own the houses and this happens through a range of different data practices um that reinforce the fact that the land is made of property. So most um maps will have parcel information on it that the parcel shapes are freely available, but you're not allowed to know who owns those maps or who owns those lands. Um the second one really became consolidated. Oh, sorry. This is them. We ended up making a map of the commas they presented to the city of Oakland. They didn't realize they're on City of Oakland land. Um, so we ended up making our own map and doing our own data site analysis um, and presenting a proposal to the city of Oakland um, before there was a major fire underneath the I880 which stopped all traffic, stopped logistics and they ended up all being swept out of the site and moved on elsewhere. Um, so project number two very quickly um, Sagoret Land Trust. So this is a collaboration with the Sigor Tra Sagoret land trust based down in Oakland which is an urban indigenous uh women led land trust that focuses on rematry the land or um undertaking forms of land stewardship and they very simply wanted a map of the Bay Area um and we ended up making them um a a map but in this process a whole range of different processes or problems were thrown up um and for this particular one it wasn't just the question of the cardographic questions of projection orientation but data sovereignty. They didn't want any of the data that we collected for them or from them to exist on university servers. And this was happening at about the same time that um one of the deans down at San Jose State University welcomed everyone back from CO with an alone skull sitting in the back of her uh camera. And so they were like, "We don't want all the shell mound data." So we triangulated the shell mounds. We actually have a a data set of the locations of all the shell mounds. like we don't want any of this data on university servers and I was like okay um and so I end up working with um our IT guy in the department I said can we have like a secret server in the university that we can ex that we can access but no one else will know is there and I don't know if some people come and they want this information I can step on it and it will you know all the information we disappeared and he's like oh I don't know about that and then he came back a few days later he's like no we do do this we do it for the department of defense and I was like excellent pre-existing policy sounds great and I think this is another set of questions about the nature of vulnerability as well um and that's being we're sort of working through this paper at the moment trying to think about what vulnerability looks like from an institutional perspective um and also the way in which it might be possible to leverage pre-existing structures as modes of resistance uh for indigenous knowledge um and the importance of indigenous data sovereignty Okay. And here's the map. I think it's a bit slow. There's a lot of data on it. No. And I'm not sure why it's not happening. Let's have a look. H. It's not working. But if you look up, the map is called Before You Are Here. And if you Google it, it's everywhere. It was launched down at the Oakland Museum of California. Um and it is on display in the natural um the gallery of natural sciences in their you here collection. Okay. Onto some more contemporary work. Carttopia. Um, so this is a project that I've been undertaking for my next book, trying to think through machine learning. And it's based from a a article that a colleague of mine, Boja, wrote, where he trained neural networks on a range of different satellite um imagery and then got those um adversarial networks to create new imagery in which he argued that we might have a security risk from so-called fake geographies. Um, and for me as someone who is um, educated in critical geography, my question was, well, aren't all geographies kind of fake in the first place, um, they're all abstracted. And so, not being able to access the archives for the book that I wanted to write, um, I ended up playing around with some neural networks and trying to think about what happens when you actually feed a fake geography, a ctography to a neural network and ask them to create more maps. Um and what you get is actually a really range of very interesting questions. Um so drawing from Katherine Mkhitri, what we know is that algorithms tell us what we already know just in the future. So there's a kind of projected element to the algorithm that actually allows us to trace across space and time. And so drawing from Watkins Fisher, the idea that we need a political aesthetic strategy, um I used neural networks. I'm in the process of using general networks to actually produce images and try and understand what AI does to an image and how it understands data knowing that we are able with many of these map collections to already understand what that data is and where it comes from. So for instance um we'll ignore bore hairs. Um, so this is Library of Congress panoramic collection. Um, so this one is a fairly small collection. It has only about 1,700 maps in it. Yep. Oh, thank you. That whilst you're talking um and what's particularly interesting is that when this um collection was used to generate new maps, there was a whole range of interesting biases towards water features, harbors, flat landscapes, and mild hills. If you do an analysis on the actual data set itself in the locations, you can actually see forms of I suppose um let's say a form of politics emerging based upon where those representations were made from. So the Midwest for instance have his fe have his features heavily in the production of those images as do a certain number a very specific group of um cgraphers. And what they would do is they would go around to certain towns who would pay a lot of money to get a panoramic map made for them which would often feature in the text boxes like the more fancy houses or whatever. And then once they're in a map in a town with a map um they would then go around and speak to all the other towns in the surrounding um area and ask if they too wanted maps. And so what you end up with is a data set that is actually particularly clustered around um 19th century expansionist policies and their representations. Um and so when you produce the images, those forms of expansionism, the plains because this is happening around the same time that expansion is being encouraged um in the west as well as industrialization. So industrial towns in Pennsylvania, you see these replicated in the data set itself. Um and so it's possible to track that using the metadata as well. Similarly, so here's a a whole bunch of different Oh, also I'm not trying to get them to look beautiful and perfect. This is kind of the key. I don't want to waste the energy on trying to make something that looks amazing. The point of this book is process to show how these images form, not to make them sort of uncannily similar. I thought about it, but I just couldn't bring myself um to retrain again over and over and over again and take um those resources from the earth to make a point that I think can be made sort of fairly easily in some of these very peculiar images and stuff. So you can see the confluence of rivers, the confluence of um industrial towns like these ones here. Um you can see port cities um and then planes here. So you can sort of see it everywhere. Um similarly it's actually possible with a lot of geologic maps. I've been working with a geomorphologist in my department U Kurt Cuffy to actually do analysis on the landscapes. Um and what's particularly interesting is the way in which um GANs normalize landscapes. So landscapes are not normal. They're not consistent. But when you either use geologic data sets or you use digital elevation models which are 3D sort of black and white depictions, you end up with landscapes that are weirdly eerily just regular and then with deep and I should have brought up I actually got a 3D model of one of those landscapes and it's really strange. Um, and I think what's particularly interesting is that even though these landscapes aren't real, and this is the point of the cardtopia, it's a place that doesn't really exist, it is possible to do forms of analysis on there. Um, and so this idea that science somehow has a ground truth um is a curious one in this particular context because um it's possible just to exist only within that data space. I'm sort of working with a bunch of theoretical ideas and this is one of my favorite ones. Sandborn. Oh, I hated sandborn. Um, it overfit again and again. It didn't matter what um sort of different kind of model I ran. Um, and this was a big data set as well. Um, it was like 32,000 maps. It would always overfit. Um, but an analysis of both the Sandborn model, which was a standardized mapping process, so it became more and more standardized across time. Um, but it also reflected the 19th century city, which was heavily gritted. And then often because fire insurance was such a big issue during those times, those maps then were used to redesign cities, the old parts after they were burned down. And so they would use the regularity of where you place fire hydrants. So those maps were actually used as information to then redraw cities. And so you've actually got an overfitting process happening in the 19th century before you even get to this complete mode collapse um that had me crying for an entire summer um before I just decided to write about the mode collapse. I'm like you know what I think this is them not me. Um okay very quickly next project. Oh there's also ordinance survey as well. Um but I can talk about that another time. Ooh. So, the final one um that I'm working on at the moment is a geospatial sea of islands with a range of students including Sophia Perez, Elizabeth Fisk, Zack Thorp who's here um and also Richard Villa Gomez. And this is a problem we just basically wanted to make a map of where films have been made in the Pacific. But it's actually not very easy to geocode locations in the Pacific. Reason being um is that actually the way in which the Pacific is imagined um both by people who describe the Pacific um but also by different national entities results in a completely uneven terrain um that kind of reflects the geopolitical mess the Pacific is currently um and so we've sort of been pulling from um Epil Hower's seinal writing uh Sea of Islands to try and think about how we can restructure the Pacific data set to more accurately reflect what it means to live in the Pacific. um and to be used by different organizations and NOS's uh for different forms of data analysis from cultural through to environmental militarization etc. So what's the issue? This is the issue got two big issues is that there are different administrative categories of islands and they're not commensurate. So the Tahesian islands will not have the same data categories as the US islands, will not have the same as the Australian islands, will not have the same as Okinawa in Japan. And even then amongst that, those islands themselves will often be different. Sometimes the Mariana Islands exists as a whole commonwealth. Sometimes Wam is separate. Um that's like the first issue. The second issue is just the sheer number of islands in the Pacific, many of which don't even have any names. Um, so we ended up exploding the USGS data set of all the individual islands. And so there's 269,391 individual islands in the Pacific. It's a lot of islands. Um, and the result and so this is kind of what the different administrative regions of the Pacific kind of looks like um from a geopolitical sense. So these are all the different um entities that have responsibility for mapping those islands. And so you can see the Pacific is actually more like a stained glass window than a consistent whole identity. Um even though we have Pacific Islander as an identity um within formal categories in the US. And then this is the second problem which is so here you can see District of Columbia um has a certain like levels of information. Um so whereas you know you can see these admin levels so county, civil township, city or town. So Washington is a city or town. American Samoa is a municipality, a county and a village and also a municipality as an unorganized atal. Um, Ham is a municipality as a village and the Northern Marana Islands are an island and an island group. Even though Ham is also both um, an island as well. Um, and even though American Samoa is not a village necessarily. Um, and so when you try and join all these data sets together, it just becomes completely messy. The only people who care about this appear to be the geologists who don't care about geopolitical boundaries whatsoever. And so we ended up getting their data set, splitting it up, but they actually um organize it all by plate, what continental plate it's on, which is of no interest to your average Pacific Islander. Um, and so at the moment we're trying to basically create a server and build a tool that will allow people to basically select islands or island groups, compile them and download them as a sort of section. And so if you want to trace the appearance of the US military from Okinawa through to warm through to Tinian through to American Samoa, you can do that geospatial analysis without having to deal with four or five different data sets that are completely incommensurate and do all the necessary spatial joins. Um, and so I think there's never been a probably a more clear demonstration of um, the problem that um, however describes where there's a gulf of difference between viewing the Pacific as islands in a far sea and a sea of islands uh, as well. island. So, we're trying to think about restructuring that um and once we've got the um actual islands themselves, then that's going to be a much bigger project as well with names cuz every island has about four or five different versions of the name as well, depending on the colonial significance, depending on spelling. So, Palao, Balao, um for instance. So, we're actually going to end up hopefully getting money, fingers crossed, who knows, to go into the Pacific and actually ask people what name should this place be um by default. And then what names would you like to be able to also optionally select if necessary? So, this is my work at the moment. Um I think I'm okay for time. So, it's not even 30 minutes yet. Um so, questions. Sorry, I know that was rather quick. Yeah. Yeah. Should we give you a round of applause? You finished? Oh, yeah. I'm done. Yeah. Thank you. Thank you. I forgot the the end at the end. Yeah. Yeah. Um I just have to run around with the microphone because of the folks online so that they can hear. But yes, we've got lots of time for questions and so many um at least I have so many. Who would like to get us started? Sorry, I've just kind of thrown this all at you. Yeah, I'll I'll ask the really obvious one that just follows up just whilst people are thinking. Um, how logistically would you go to all of these thousands of islands and ask people what they should be called? Tell us a little bit more about that plan and I hope you do get the funding for it. Um, so I mean this is a sociologist project really. Um, can't go to all thousands of them. Most of them are uninhabited. Um, there are already groups working in Pacific working in geospatial data. Um so there's the um Pacific Ocean Geospatial Consortium for instance and that's a place that I would start. Um and then it would be just a question of actually reaching out and building a multinational project with people on the ground who are already doing data science. Like every place has someone working in data um and it would just take time. I mean the map that I couldn't show you that I work with Sagor that took 3 years. I I was drawing hills on an irregular stereographic projection from three different perspectives with no top for 8 months. Handdrafting actually I was sharpening um mechanical pencils on um sandpaper in order to get the points that I needed because it just this work takes time. Um, and I think that's part of the theoretical question as well is that you kind of end up going against the grain of what's accepted um, if you if you ask these questions. But the reason people don't do it is because it takes time. Um, but I'm an academic. I've got time. I'm not an activist. So I sort of see my job as if no one's bothered to ask, my job is to spend 10 years asking um, and trying to make that happen. Amazing. Does anyone have questions? Yes. Um, I'll get you started, Heather, and then Gayen. Is that okay? Heather Hman, um, sociology, the business school, bids, many other things. Um, so I'm I'm thinking about this uh, a sea of islands. So, I read something recently in either The Atlantic, The New Yorker, or the New York Times magazine. They blurred my mind, so I have no idea what it was. It was not The Economist, though. I know that. Um, about the um, sorry, I should be like I'm kissing it. I know. um about all the currents and not knowing the sea itself and are there geographers and ctographers working on currents and you know below shell I mean that you can map the shelves and stuff but trying to figure out like it's the the sea itself also needs to be mapped right in ways that make sense to the people living there and know they know make sense of it because it's their livelihood. Mhm. So this is the sort of holding map that we've made for the film database. Um the first thing we identified was there is an issue as well cardographically speaking not data but when you zoom out for the Pacific it's just one big blue blob. Um and so we're thinking about actually how do people know and understand the stories of the Pacific. And I am extremely fortunate that one of um the graduate students in my studio, Sophia Perez is Chamora and she is learning how to sail from the Carolinian navigators who are the descendants of um Papa Mal Puig who was the navigator who taught the um Hawaiian Renaissance people um on the Hoka Hoka to sail using traditional techniques from Hawaii to Tahiti. there is a bigger project which is us like we can get the surface current data um and we can triangulate it and we can animate it um to some degree there's a lot of um and yes there are a lot of money someone. Um so yes, there are I think the center for ocean as well. Sorry, I think your I think your battery um on the microphone has just died. So the folks online can't hear you. So you just do a very short sorry. Um yes so there are people working on it. Um and I think this is kind of the interesting thing about these ways in which we separate out data like oceanographic data is different to environmental data is different to logistical data is different to social or cultural data but actually they are kind of related. Um and I think what's particularly interesting I'm not sure that you can see it at this resolution. No, not so much. But for Sophia says um that actually the currents mean that it's very quick to go from Saipan to Tinian. It only takes a few hours, but when you go back because you're moving against the currents, it takes more than twice as long because you've actually got to triangulate like this. Um and so that is something that we're thinking about in a cardographic sense as well. Fascinating, Gayen. Thank you. Um hi, my name is Galen. Um, I'm a maintainer of uh the Arches project, which we are an open- source GIS CMS platform, and I'd love to talk to you about it after this. Um, I've got way too many questions, but I'll limit it to two. Uh, one is about uh projection, geospatial projection, and that as a sort of colonial construct, which you touched on at the beginning. Um, you know, I'm sure that like I'm sure that there are conversations about like what are we moving towards? What are what are Yes. Yes. Thank you. Yeah. Uh I guess I like I think about like the Peterson projection as just sort of like one of the earlier examples of well this is how it could be right and how Antarctica is not accurately reflected in WebMater. Um sorry esoteric. I I'm curious I'm curious about uh what you can tell us about that momentum towards what you just pointed to like decolonial projections. Um and then the other question that I have is more about um uh apppropo data sovereignty. Uh for example, you mentioned how you wanted to be able to turn off the server uh that was holding indigenous data. Uh I'm wondering if you could tell us about for example if there are decentralized data solutions that you looked at uh for the purpose of data sovereignty especially for tribal data. Um yeah thank you. Um okay so projection. Yes, big problem, cardographic problem. Um, so the best example, so this is before you are here. So this is a map that um I made with Sigorite. It was supposed to be a digital map. It still will be a digital map. Um, but the digital map got delayed by 3 years when I said to them, so I can't make a map with an indigenous organization without talking about Mada and the Greenwich Meridian. And they said, what's that? And so I explained to them how in 1886 a bunch of um honestly white European men kind of got together and decided that the Royal Observatory um in Greenwich would be the the through line um for the longitudinal datim. And they went, "Oh, no, no, no, that's not great." I'm like, "Okay, well that's fine. We can repro. Where do you want 00 to be?" and they said, um, we want it to be down at Lashon, which is one of the first sites they had rematriated. And then I was like, okay, cool. So, we've got 00 sorted out. What kind of projection do we want? Which way do you want to look? What's going to be at the top? And they said, why does there have to be a top? Uh, why should we only look from one way? And I'm like, these are very good questions. um because enlightenment thinking, because perspective, because ge, you know, geometry, but actually none of these are really reasons. Um and so in the end, we have a map that you can turn around that has no specific top even though it has a top, so you can turn it. And the only exception is the Tuli, which is down um in South Bay, which I could not make look correct no matter which way it was. Um but you can turn it around so you can view it from any angle and it's got a kind of gestalt as well. So technically so what we did is we created 00 then we used a tool called the relational reprojrojection platform that was made by two Berkeley alum will pay and Eve Mclling from Torston Hagarandran's relational geography and that tool actually stretches out and reshapes the area. So we reshaped the projection into an irregular stereographic projection. So it's a set of rings that are at different distances. Um, and then we did a really high-tech job of actually putting a base map on a projector and putting a paper on a wall and moving the projector was on a trolley and moving the trolley and matching it up and drawing. There's a great picture of even I like just drawing. Um, and then I sat down and spent three years getting the original close coastline in hand. um sort of using old school drafting tools to make sure that that was um in aspect to this irregular projection. The same with the hills. So I had like a digital elevation model and a whole bunch of archival images that were pasted around the entire outside of my drafting table um that we would turn and draw um and then the stars. So So you're looking down at the coastline, across at the hills, and then up at the stars. And the stars are if you were standing at Lashawn looking over Mount Diablo or Tushtuk on the eve of the winter solstice in 1578, the year before Sir Francis Drake kind of sailed and put um California or sort of the Bay Area on a map, that is the constellation you would see. Um so if you sort of turn the map around, that's how you would see it. Um so yes, projection is a major problem. How are we going to do this digitally? We don't know. Um, I've got a colleague Nick Lai who's been messing around with different tools um to try and make I think 3DJS is possibly an exciting moment. And then the second one was data sovereignty. No, I haven't thought about decentralized systems mostly because I am not a great networker. Like networking um is not my strong suit. And I have a colleague who's at RMIT University in Melbourne um indigenous colleague and she's doing her PhD looking at the possibility of using blockchain for managing country. And so I'm sort of waiting to see the results of her PhD um before I move into that. Also honestly to preserve myself because I have tried to understand blockchain and I don't feel like I can adequately explain to anyone what the risks and benefits are of using that kind of decentralized system. Um, so at the moment it's heavily centralized and I have promised Sagoret that I will step on the data and destroy it all if the university gets taken over by, you know, uncool crews who might want to harm them and eventually that server will be transferred down to them. They own it. I bought it for them. It's got their name on it. Um, it is their property. That's so that's so wonderful. Um, Fernando, let me run the mic over to you. Thank you Christie. Uh thanks for this wonderful talk. I'm Fernando. I'm faculty in statistics and kind of connected to to bids as well and and other efforts on campus and I'm I'm asking you've covered a lot of wonderful topics. I'm going to index on one teeny bit that I happen to be very plugged in which is the tools part. So, I'm I'm a bit of a tool builder and I'm curious what what do you see existing and what do you see missing in what the tools you would like to have in the kind of open ecosystem for this type of work and how could perhaps places like bids and others on campus help you and assist you? Yeah, I mean at the very base level I've been talking about this trying to figure out how to get money to build it. Ultimately, I should not have to be doing so much work to represent worlds on behalf of people who should be able to represent them themselves. The biggest problem is open- source JS like Q down here is clunky. It's really awful. Um, once you get used to it, it gets better. But as an opening thing, you open it up and you're like, what am I even looking at? What do these logos mean? It's it's really offputting. ArcGIS online is better, but it's ArcGIS. Um, I would love to eventually have some kind of version of ArcGIS online where people could choose where their servers were located or even have a server located in the sovereign territory of of an indigenous nation where people could start by saying, "I want my 00 to be here. I want to use this projection. I don't want all of these other roads and building stuff on it. I want to just put some points on it and I want to share it um in the first instance and then over time. So uh a colleague of mine Nick Lai um has a really cool tool that he's building out um called shaping. It's web- based and it does this kind of work. So at the moment it's raster and you download it into vector. But the idea that eventually you could also have people create their own maps that are not just a consistent cartisian grid where all spaces are theoretically a sort of equal but they could actually make some of the spaces bigger and some of the spaces smaller which is what we ended up doing by hand with the Sigora Tay map. You know they wanted their home to be bigger. They didn't care that like theoretically speaking the UC campus is geographically bigger. It's not emotionally bigger for them. Um so that people can actually make spaces bigger and smaller as well. Like that would be the dream. Um and then you know actually again paying attention to data sovereignty in a way that ARGIS just doesn't. Like I can't ask people to put information on a server that's you know shared. So that would be amazing. um just out of the box web- based and that comes from as well a piece of research I did about teaching. Web- based is the most accessible way that people can access information. Everyone's got a browser. Um ARJS doesn't work on Apple. Um ARJS it doesn't work on Apple computers. Um Q doesn't really work on Chromebooks. Like it's a total nightmare. But the browser, okay. Um but then thinking about that sovereignty. So that would be my dream. So I I all need to talk. I'm gonna run the microphone over to Tim, but there are literally emails from Fernando over the weekend that are hitting exactly these challenges because you're not alone in feeling them. But the way that you're articulating the need is absolutely beautiful because it pulls these different these different challenges. And I can I can bring you people like we have I've got a kind of there's about four or five geographers who work with like the CRE and different tribes who are just like we just need something like this um that we can ethically say yes um we can do it and then we don't have to retreat into our own institutions and institutional knowledge because I mean we end up having to protect them from the institution as well. I mean, I will, but some people can't. Um, and particularly in this climate, you know, we're even debating whether or not we want to keep our service here or move them down to Oakland. So, uh, this may be a little ignorant and more conversation starty, but um, uh, what about like, uh, current tools like, uh, Map Libre? Yeah. And Kepler. Yeah. Where are they at? I know I mean Kepler's Uber which is problematic by and large but like is Map Libre open source enough to get us a couple steps towards your vision. Yeah, it's opensource. I think there are there are enough open- sourceish kind of things out there. The problem is the user end like actually just being able to access it, get on and make a and I I had honestly hoped that felt the mapping platform would sort of maybe do something like that, but it commercialized extremely quickly. Um, and you can't choose where you put your data. Um, so yes, there are options. Um, but again, just that ability to actually be able to go in. And I think something interesting that Karina Gold said to me is that her and a group of other people dragged into Google for an indigenous mapping workshop. And she said, it always stuck with me. She's like, so they dragged us in all the way down to South Bay and taught us how to put some points on a map, but what for? I'm like, well, this is the thing. What for? Um, and and this is the problem, which is that you either got the very simple what for? or you I mean even map Libra you need to get into the back end into the code you need to know how to script can I ask a followup to that is the university a space to avoid selling out or whatever but is I mean obviously right now it's very difficult we don't know where it's we're going but to create an open- source almost like nonprofit corporate to create these tools I think the university is one of the few spaces left um because it has enough cash and institutional power that it can take the time that it needs to and I wouldn't say the university as a structured all the time. One of the reasons I ended up here is because we ended up we made a situationist inspired um app designed for field research and the university that um we had the grant at which was a European grant kept wanted us to commercialize it essentially and said we went we did not own a source code and we had to commercialize it and we had to commercialize it by selling the data. Um and we were just like um this is the app is like this is situationist that is not in all it keeping and so we lost the source code um and the app was never published and never used and so um this is one of the reasons but I think here we do have a lot more control but I do think agreements with organizations that we're working with that protect them that guarantee their sovereignty their intellectual rights and stuff is really important just to make sure that that protect protection exists. Um, that is a truly wonderful moment to finish on. We've got just two minutes and I'm sure everyone has other classes and and meetings and talks and things like that to go to. Um, could you all please join me in thanking Quincy so much? Thank you.