You're listening to episode 751 of A Very Spatial Podcast, December 1st, 2024. Hello, and welcome to A Very Spatial Podcast. I'm Jesse.
I'm Sue.
I'm Barb.
And this is Frank.
And this week, we're going to tell you some things about geography. Imagine that. Mostly news items, though. And one that we begin with is, is one that is not new. It is a repeat from 2022. Oddly enough, Microsoft is highlighting Microsoft places again. I'm not sure. If it's because people didn't notice it, maybe they're telling it and people are saying that it's new because they're adding AI.
Yes, it's absolutely because nobody noticed it. It really because nobody noticed it.
It's an add on. All you have to do is go into the add ons. It'll tell you that it might cost you more. Or if you already have teams premium at your organization, you already have it. And if you don't have teams premium, then potentially, if you have an exchange license, you may have parts of it. Like most, Enterprise software, you just don't really know you can try adding it to your account and see if it works
and and to be perfectly honest, I don't I don't I still don't get the point. We had a conversation about this already, right? Go back and listen to episode, whatever the hell it was. And I still don't get the point of this thing. I just don't.
So you have these rooms, right? And these rooms aren't static rooms. They're able to be used by anybody. And then you have people, some are in the office, some aren't in the office. And the software goes through and says, okay, well, you want to meet this room is available this time. And you can. Have this many people in that room. So use that room with these people and everybody, you have this now added to your calendar or something like that. It's, it's not really places like, but
it's no, it's a terrible name. Let's start with that. But first of all, there's tons of things to do this already.
And second of all it talks about the news article we linked to it talked about like hybrid work and remote work and stuff like that and not to get into a really tangentially geographically related sub rant but you know there's all the stuff in the news now about how a lot of corporations are canceling these things and they're moving towards a more traditional everybody in the office type of scenario and it's very interesting to me that the corporate world
is Still wrestling with the notion of disjointed geographies in the workforce, but coming up with solutions, software solutions to deal with this thing that they don't know what to do with. So they get rid of it. It's just a weird place for us. No pun intended for us to exist.
But it still works for if you have everybody on site as well. So,
yes, it has like a was a nearby function. So. This way you can't hide, I guess. So they can know that, that that only is so and so in the building, but guess what, they're like two rooms down the hall.
That's not creepy.
Or whatever it is, so. And they're also going to add co pilot to it, I guess, later this year to, I don't know, I'm not sure what co pilot could be. Can add to it too much. But hey, I don't, I don't use it. So
I don't know, but all I was thinking is this would be a great time for them to bring back second life and then add a second life function to it. So you've got the rooms, but you've got your avatars and they live in that meeting room.
Well, there's tons of those out there.
Yeah.
And you can even do them with head mounted display. So if Microsoft hadn't just killed off the HoloLens in their VR Focus areas. Actually, I'm not sure how much they killed off, but the actual Microsoft partnered headsets also been discontinued. But I think they're still supporting other either way. You can still do, you know, virtual rooms in addition to virtual rooms.
So here's here's my creepy story of the week writing. This is where I think the copilot thing just makes it creepy is, and this is car play has nothing to do with Microsoft or anything that they're offering. And it's not work related whatsoever. Is that okay? On Monday nights, I take. Guitar lessons at a local music store called Follies and clearly my phone has realized that a certain time of day I go to work at a certain time of day.
I go home and when I open up CarPlay in my car by the default screen, it has my music and it also has the navigation screen and it usually shows me this is where you're going to go to next. Well, the other day I got in like on Monday, I got into my car, And it went, okay, here's how you get the follies. And I was like, well, first of all, I know how to get the follies. Second of all, it's kind of creepy that, you know, that that's where I'm going right now, cause Most weeks I do it.
Sometimes I'm busy with work or something and I just can't make it to class. That's just creepy. I don't even, I can't even imagine how creepy it would be for my team's app to tell me, Hey, Bob's next door and he's free for the next hour. So you should go harass him. I'm like, just leave me and Bob alone. Okay. We can have a conversation if that's what we want to do. That's all I'm saying.
One of the things that I teach when I teach basically what would be considered the digital earth class is we start out with the psychology of creepy. The geography and psychology of creepy and go over it because then we get into the timelines of geospatial technologies. We can talk about these things about, you know, thinking about how people receive them and it's always really interesting to go through and think about, you know, how do people feel about this technology today?
How did they feel about it in the past? And, and what's the creepiness factor, which is it, it knows more about you than you want it to
continue on with a creepy factor. If you have been using Foursquare and nobody to be honest is, is probably checking in anymore on Foursquare, but if you have in the past, your data has been now used to help create the Foursquare open source places data set. And this is a base layer of 100 M M plus what, what am I supposed to try to take with 100 million million plus? Yeah. Yeah, there are two M's there.
I'm gonna, I'm gonna assume that it's a hundred million, over a hundred million global places of interest which includes 22 core attributes. And it's updated fairly regularly. Yes, please.
Square meters? I mean, technically two M's would be squared, but you know, a hundred square meters is not that much. That's
The places you would think would be point data, so it would be quantity.
Yeah, I mean That's that's interesting. We're
still
a lot about that. It's a lot.
Yeah. And they'll tell you more about it later on in the blog article. If you go to the show notes and read it. But I, I think the key here is that, you know, there's a recognition that four square has not been pushing their data. On a broad level, they have been, you know, working with commercial organizations and such and selling it in various ways. But this is a recognition that between and they even mentioned things like open street map.
And so looking at various open initiatives, like a street map, of course, newer initiatives that are building from kind of this corporate enterprise level. Of data collection and up to make unified data sets that lots of people can access. So there is this trend to, to try to support this and that's again what Foursquare seems to be doing with this release of open source places is that now they're providing this data set is kind of a foundational data set.
That can work in conjunction with others for you to use in, in these different ways. And they talk about it as a cohesive and comprehensive place operating system. A. I contributed at scale confirmed by human contribution at scale. So, you know, they're looking at how they put it together and how they're going to take from here. And you can look to see what they're looking at now. In the show notes as well as their place makers. That's the name, right? Yeah. Fsq place maker position.
So they're releasing these now, and they're going to open it up for other people to begin to utilize it and grow it. Because of course, this isn't meant to be a static data set. And so you can go and find out more about the place maker part of it, which is coming in the future.
So you did discover a typo. I'm pretty sure because when you go to the schema, which is linked in the article that we linked in the show notes, and you can find a lot more about what kind of attributes and things that they collect. It's kind of interesting, particularly if you're a schema nerd. Then you go to the about piece and look at it. It's 100 million places. And I think it's supposed to be 100 million plus places because it is slightly over 100 million places.
And I think somebody just inadvertently typed an M twice.
Yeah, thoughts on Foursquare moving into the open source places?
Well, I thought it was interesting too that, that in their, Post announcing it that they kind of went back and said, look, we've benefited a lot from open source over the years and utilize that stuff and and recognize the importance of having having those things out there. And that data is one of the things that has. Has often been proprietary.
And you see the cool stuff people, you know, they kind of guard it closely So, I mean, I think this is it's gonna be a really interesting data set And you can you can get the current or the latest version of the os places now It's of course the placemaker where you get to contribute that isn't quite ready yet and you can kind of join a list to get notified on that But I mean, I think I think there's a lot of A lot of potential here. In fact, I may may download it and take a look at it myself.
I thought it was interesting that one of the reasons why they're doing this is they talked about the maintenance of such a large, real time, dynamic data set that, you know, that that part of doing something like this and keeping it alive and keeping it useful is having to to maintain it and that their, their hope is that by making it open, that will make the maintenance. easier, faster, more efficient.
It is worth noting that if you go look at the places API and you go, well, it's not quite everything I wanted, even though it's free and it's open. That's cool. But they do have something called places pro, which is a premium data set you could subscribe to. And it looks like it's using machine learning models like everybody is now to try to, you know, extrapolate this data using other data sources or link it to other data sources. So there are them. You know, subscription options available.
It's all an early access for right now, according to the page. But, you know, they are still continuing with some form of pay model. If that's something that your organization needs or prefers,
you know, we can point to organizations like Overture maps as kind of a comparison, though. Of course, there you have a lot of different organizations that are working together, but. Foursquare is not one of those. So while you have organizations like Amazon, Meta, Microsoft, Tom, Tom, Esri, and the list goes on, it does not include Foursquare. So, so kind of a I, I would think of it as working with both things as opposed to one versus the other.
I was going to say, you just take them all if you need them.
Yeah. Open is open. You know, I got to be honest with you.
I haven't thought about Foursquare in a long time. Since the last time you checked in with this four square app, I, yeah, I mean, I think it's probably five years, you know, which is, it's interesting because it was such a thing for a while. Right? And so was, for example, Pokemon go, they're both kind of things that sort of had a shining. Star that just sort of Wayne
fair, Pokemon go is still in the, if not ascendant, still stable.
Well, it's definitely stable. I'm just saying that, you know, it, it just kind of went into the popular culture and then kind of went out. And I think that an interesting thing, I genuinely believe that the COVID lockdown situation is what partially made that happen. Like you just don't. You didn't think about that stuff anymore. Cause I was just thinking, when's the last time I thought about four square? I was like, well, it was before COVID.
So when you really couldn't go anywhere to do those types of things anymore, you just kind of forgot about them.
Of course, speaking of Pokemon go, Niantic is our next organization we're talking about and their attempt to, to utilize again, the users generated data to create something. And so talk about LGM. So large geospatial models.
As opposed to trying to figure out how to get LLMs involved and working in the geospatial arena and geo AI, they're thinking of this as the foundational model being the geospatial and they're using it as part of their already existing visual positioning system that they've had people, you know, scanning. All kinds of things that are pokestops and such around the world for, I don't know, that's been in the game for at least half the game for the least the last five years or so.
And so, yeah, it's just, it's interesting to see that we keep seeing companies coming around to, well, you know, we're, we're. Doing a lot of data collection, and this is finally a way that we're going to make it possible for more people to use. So whether it's in Pokemon Go or one of the other five games that I think still are using that same foundational code, they are still collecting data.
And if you go to the show notes, it gives you kind of a map of places where they Do you have lots of things mapped? And so using these small scale you know, sorry, large scale, small area data collection figuring out how to tie that into broader data sets to go from this local to the global. It's kind of what they're talking about in this particular article.
Something that is interesting in the article that they jumped out is I'm just gonna read this when you look at a familiar type of structure, whether it's a church, a statue or a town square, it's fairly easy to imagine what it might look like from other angles. Even if you haven't seen it from all angles as humans, we have spatial understanding, which means we fill in the details. Blah, blah, blah.
It's interesting to me because I was what immediately popped in my head when I read that was how I can't even begin to quantify the number of times I've been really wrong. You know, you just go to the church, you have this and you go to the back of the church and go, oh, wait, it's got a whole thing out back there. I didn't even imagine was possible. And so I think this is a really interesting approach. And it's also going to be really challenging.
And I'm wondering if it's going to end up being a little bit like generative AI of art. You know, you make the thing that looks like what you think it will be. But humans, most humans don't have six fingers, you know, that type of situation where it infers something there that just isn't normal. It will be really interesting to see how this turns out. I think it's going to, I think it's going to be an iterative, well, I almost guarantee it's going to be an iterative process.
It's going to take a few years to work out a very Functionality for the general public
and I actually zeroed on that part too, because it made me think back to some technology that never made it out into the consumer world, right? But that we use and talked about on the podcast to pass. I could stop photo sent Microsoft photo sent and we all thought that was really cool, but that was the weakness of it, right? Is because it scraped the Internet.
To create, you know, 3D point clouds of things, but when we tested it and we even tested it in our virtual reality cave at WVU, is that the biggest problem was, is that you would not get the pieces that were not of structures of landscapes that were not in the photographs, right? People tend to face it and it's their view. So. It's kind of, you know, and it's interesting. So that was a huge weakness of it.
We, on our small attempts try to, to kind of mitigate that and using the point clouds to create things. So, so a thorny problem that that's been around, right, is recreating the world when the humans that are doing it can't go certain places or don't and then, you know, you have to kind of, as I say, extrapolate if you want to have that three dimensional build. So it's, it's, it's, An interesting, interesting problem, but I, but I think you're absolutely right. Frank did that.
You know, I mean, you assume for the purposes of these models, right? The sameness of things, a church will share characteristics, but the reality is that even at large scales, there's so much variation, but then you also say, well, how, how, how many people will ground truth that? And if they do, you can update it. So, so I think it's interesting.
Yeah, I was thinking about the ground truthing because it made me think of an interview that was with the founder of OpenStreetMap, where he talked about apps like OpenStreetMap as the reason you get so much data is because it's almost like a hobby. So it made me think that all these things that people do for fun are going to be really powerful for ground truthing because they're more likely, you're more likely to get. Input or multiple inputs on a location.
Yeah, I think that the interesting thing here is that machines will always struggle with the just like humans. We struggle with the two. We make a presumption of this is what should be next, but we don't have all the information is incomplete information. So. There should be an X, Y, Z on the back end of this thing, and it should look like this. And then when you go look at it you go, well, it doesn't look anything like that.
And then you find out because somebody made a big donation of a ton of bricks, literally that, Oh, that's why they built this whole wing on the back end. Or they ran out of material and they shortened that a little bit or, you know, whatever it may be that type of who would know that. And then a lot of that's lost to history too. So, you know, things that we have to do and, and really it's the Area of expertise more than mine, but you have to sort of infer and Jesse's infer things based.
And there's a lot of times it's inferred based upon current norms and understandings and all that sort of thing. And in reality is you may never know exactly why the thing isn't at 90 degrees or whatever that you're. Presuming must be true, but factually isn't and I think it's going to be very difficult for the large language, large language, large geography model to incorporate that fuzzy information.
I think the real power here comes when the large language model approach concept thought, I don't know what the word I'm looking for here philosophy integrates with a large geography model to get that more fuzziness in there and be able to have that degree of, oh, well. This was made by this builder and he had a perchance for making things at, you know, perfectly right angles or not perfectly right angles or whatever may be.
But you can also see, of course, where we can take the the collected footprints from an area, which we can do through aerial activities and machine learning, and then tie that in to the large scale cloud point cloud data that, you know, groups like Niantic and others are collecting and being able to. Kind of invest some of that. Well, here's the footprint. What would you do with the footprint? And really, it's kind of an extension of what city engine was doing whenever as we bought them.
And I think, you know, I think we can all agree that that was the reason that they bought that product and brought it into ArcGIS because it was a way to do supervised terrain generation. So, you know, we're already talking about the use of AI and doing that there. It's just a different idea from what we're looking at currently in terms of the generative models, those things that we're playing around with. So, yeah,
yeah. But I think, and, and. This will remain to be seen, but I think what Niantic has hit on, right, is they have a million, millions of users, right, a huge user base that they've trained to, as you say, add PokeSpot PokeStops and do all of these things and collect that information.
So that's an enhancement that these other software packages and these other models need to try to generate landscapes That's an advantage that they didn't have that you can in, and again, it'll be selective, but you can essentially ground truth and train the model even better. And so that's, that's the interesting part is to see what they use that for.
Amusing thing for me as someone who regularly uses Pokemon go is they are desperate for people to use that scan function and just people aren't. Using it, because the return on investment of standing there and walking around something to scan it and make sure that it's perfect for the way that software wants it to be done is just not, you know, unless somebody's just standing there and going to be there for a while. Anyway, there's just no real incentive to do it.
So they've, they've tried to, and then they've tried to use a stick to get people to do it, and it just, they've, they've not been successful. So enough people are doing to make it useful.
Yeah, I was going to say that's the small percentage is still more than others. I have the advantage of
like a, like a geospatial modeling Ender's Game training. You're the game.
That's that's a deep cut science fiction thing for anyone who, anyway, I did want to pull out a couple of things from it
because we've never done deep cuts on science fiction.
That's true. It's never happened. I did want to point out a couple of things that I really liked in the article. 1, it's a visual positioning system. Okay, for you corporate people out there, you know, not everything has to be a positioning system. I just point that out. It's doesn't have to be. Secondly, It's 50 million neural networks. That's impressive. I mean, large is barely covers it. And then you read 150 trillion parameters, not data points, parameters.
My goodness, that's large just seems a little bit of a. A little modest.
We were all worried about Skynet. No, it's actually Niantic, right?
Yeah, that's, I mean, amazing.
And so, speaking of people that we always worry about, then we can call them Alphabet, but we'll just talk about them as Google. Google, if you are a person who has used your Google location history, so either turn that tracking on on the mobile app or use it in the browser or any way that you've, you may have used in the past, keep in mind that you are at the end.
So Google location history in the web is going away, and it may have already gone away for some people, but everybody else probably has a couple of days at least left to capture it. And it will be, I think they said available for download for a little bit longer, but you're not going to be able to actively add it. And that's because it's moving to on device.
And once it moves to on device, there'll be, I think, a short overlap of time where it's on device, but you can still download it from the Web. So if you want your Google location history, go download that. You have to go into the what's it called the take out. Yes, the takeout which of course from there, you can download any of your Google information and delete it if you want to as well. So go to Google takeout.
If you just want the location information, deselect it all, just grab the location by selecting it and doing the, the save. But yeah, I mean, it's, it's always there for you to be able to do that, except for this one, which is going to be going away.
I read this and I thought, Oh yeah, it does that. I forgot. It's just, well, even there was an option and then it was, and I'm pretty sure the reason I'm mine was never turned on was that I had the exact same thought pattern in 2017 or whenever was available for iOS, I went, Oh, that's interesting. How would I use that? I don't know. I'm not gonna turn it. I always suspected that was a very dominant pattern, which is why I think you would find.
I mean, obviously, they're moving it toward device for a whole lot of security reasons and things like this. And I'm, I can't remember any of the numerous lawsuits that. Alphabet or Google has suffered in the EU change business practices. This may be an outcome of one of those for all I know.
And many times companies like this find it just more advantageous to just blow it out everywhere rather than just say, well, if you're in the EU, it works this way, if you're in this place, it works that way. So. You can see why it's happening, but also, what do you use it for?
Well, I use it to show my class. So, this is where I went on this trip, where I went up to New York, and then West Virginia, and then went over to, after coming home, went over to San Diego for that summer in 2012.
Ironically, the one I use for my class looks exactly the same.
I wonder where you got it.
I wonder how that happened.
Yeah.
You know, other than no, your honor, he couldn't have done this, whatever bad thing, because, you know, that type of scenario. I'm like, why, why would I use that ever? I mean, for the most part, it's kind of interesting to me. And this is a thing that I've run into. This is not geospatial, but this is the thing I've run into that I think has been a bit of a challenge is that.
The way I'm used to thinking about computing technology, and we've talked a little bit about this, is that, you know, you have folders subfolders, and I don't necessarily remember what the name of the file of whatever it is I'm working on is. I just kind of remember I would have put it under here, and then I probably would have put it under here, and then that's probably where it was at, and then I could find what I'm looking for. Instead of of remembering exactly what it's called.
And I can imagine then in a timeline situation like that, where where you've been for me is like, Oh, what was that coffee shop that we went to that? I really like the coffee. Oh, it was on. This day and we went up to the strip district in Pittsburgh let's just Google all the coffee shops and strip district in Pittsburgh and then I'll go, Oh, it's like around the corner from Benzie's.
And then I go, Oh, yeah, okay, there's, that's the one that I must have really like rather than like, just go look at my map and go, well, on one, I have to remember what day it was there that I was there, but that's just my approach to thinking about location and a much more fuzzy approach as opposed to that specific. And this, I just don't know where I would use the term. This functionality, even on device
to address what you were asking earlier. Yes, some of the EU things, but this is also to keep alphabet slash Google having to respond to every organization asking for location data for person X, because they don't hold it anymore. So they can get rid of that entire section of. Of data requests, because ain't got it. Couldn't tell you. Yeah. So like, Oh, Apple did that and they didn't have to answer those requests. We're going to do that too. Yeah. Right.
But yeah, that was something we talked about. I don't know, like a year ago, whenever they first announced this, I think that's, that's kind of, is there any other news? Is no, well, then the web corner we have a web corner this week. It is one that is not new. Okay. It is one that has been sitting around since 2021. Just want to remind you of the, you are here podcast. That's the letter you, the letter R and the word here all squished together. You are here podcast.
And so this was a quick series. It was 11 episodes, I think back in 2021. And. And it was a podcast by Gretchen Peterson and Vanessa. Knoppke-Wetzel.
I'd say it's Knoppke-Wetzel. That's
Knoppke-Wetzel. Okay. We're going to go with that. So we have that they were the, the show hosts and each week that they did this back in 2021, the spring of 2021 basically they looked at three maps and talked about those three maps and yeah, it's, you know, there's lots of podcasts out there. This is one that sadly isn't currently. In production, maybe they'll, they'll bring it back because somebody was posting about it recently. I think it was Vanessa. In fact, I had something on blue sky.
That's what reminded me of it. Yeah. And maybe we'll do a few weeks of web corners at our podcast because they're unlike once upon a time when there were a couple of us out there and then just us for a very short time. Now there's many. Whenever we talk about either traditional ideas, a podcast, or getting into some of the video podcasts and YouTube channels that are out there. So, yeah, we'll do a few of those just to make sure others know other sources to go out and watch to check them out.
You are here, podcast. com and all of your favorite podcasts. Distribution sites. Okay. Then on to the events corner. We ain't got no events.
It's the end of the year. You know, things get to keep it local. People do events locally. Spend time with your friends and family and colleagues. Do that way to do an event.
I'm putting together a GeoBee from all my students creating questions, and it's a great quiet, fun activity to do with friends and family. I'm also gonna be doing a geo nerd one. A few of us that wanna get together have decided for Christmas. We're gonna create our own GeoBee, and then everyone in geography's gonna play it from wherever we are.
If you like us to add your event for the next year, end of the podcast, send us an email to podcast@veryspatial.com.
If you'd like to reach us individually, I can be reached at sue@veryspatial.Com, and I am dipping a very small toe in the water at Blue Sky. So mostly just, just looking at feeds, haven't posted much yet, but you can find me at GeoGirl there.
I'm barb@veryspatial.com
I'm Frank@veryspatial.Com, and I've gone a little more let's just say knee deep. Into blue sky, I posted a few a bit and I'm starting to post more and it's it's a nice place. So check it out. @nojopar
and I'm available at kindaspatial. And I have after threatening it for probably a year and a half updated. The contacts at veryspatial.com/contacts, at least for the social media to reflect everybody's blue sky. And if they do have a threads, I think I also linked to threads. Let's check and see very spatial. com slash contacts. Oh, no blue sky and mastodon. I did not put threads on there. So if you, if, if you guys have, Mastodon accounts or thread accounts.
You want me to link to email them to me? I'll I'll add them to the page.
You know, it's interesting if you look at the usage of blue sky and threads I mean threads unfortunately for threads. It was just kind of dropped off big time but I do have I do have the threads. I will I will let you know As always,
we're the folks from very spatial.
Thanks for listening,
and we'll see you in a couple weeks.
You. While you have the chance Oh, why didn't you do Something relevant? And soon you'll realize why you started it all over. Oh, the flickering times of the running back, the running back. And I think it's worth something more than to realize that there's more. I think it's worth something more than to realize that there's more. I realized, oh I realized, at this moment.