Surprised to find this on HN this morning! I'm Director of Engineering at UrbanFootprint (https://urbanfootprint.com/), the company featured in the Fast Company article. We provide data and tools for urban planners to assess and compare the impacts of land use and transportation decisions.
A basic use case is a city updating its General Plan, which would start with a forecast of how much population growth is anticipated / needs to be accommodated. A planner then needs to assess where new residents will live, work, shop, and play. Perhaps even more essential, how are people going to travel between all of these activities? Will the new growth be auto-dependent, transit-focused, walkable? Is any of the existing or planned development in hazard areas such as flood of wildfire? What are the energy and water use impacts of the plans?
We’re using Python and Postgres/PostGIS on the backend to answer these questions and a React SPA to serve it up and make it interactive in a browser.
Also, if you happen to live in California, UrbanFootprint is available for free to your city through our California Civic Program (http://info.urbanfootprint.com/california-civic-program) so feel free to nudge them to get in touch ;)
Congrats for the PR successes, and also for turning what I can only assume was a worrying SimCity habit into a career & a tool for good.
To expand that SimCity analogy to its breaking point: I've long fantasised a game/simulator crossover that tries to get as close as possible to simulate a city. Down to, say, individual agent's decision to eat out or shop for groceries, and the resulting traffic etc.
Is that something you see happening in the (longish-term) future?
This is already in common use as far as I'm aware. Its used in the Netherlands to predict electrical grid load based on the number of EVs driving about.
Today, one of the biggest differentiations between UrbanFootprint and UrbanSim is that I'd call UrbanFootprint a "sketch analysis tool" whereas UrbanSim is, to the best of my understanding, an agent-based simulation tool. For a user, the key differentiator is that in UrbanFootprint you build scenarios with the actual future land use that you envision and we run analyses to provide insight on the relative impacts of those scenarios. So you might build out a few scenarios like “Business as Usual”, “Urban Infill”, and “Transit Oriented Development” and assess the cost/benefit of those different plans. In this use case you are literally “painting” land use onto a map (with the aid of lots of nice faceted filtering, geospatial joins, etc.). In the absolute simplest terms, I like to think of it as “you tell us what and where you want to build and we’ll tell you how that pencils out.”
Additionally, we aim to provide numerous high value datasets and analysis modules “out of the box” to lower the bar to entry for many planning tasks. For example, we provide users a normalized nationwide parcel-level land use “canvas” to create projects with. This means you can be up and running with parcel data for most cities in the United States in a few clicks!
And for the record, there is absolutely nothing on the technical side that precludes us from doing agent-based simulation -- we just haven’t focused on that yet. I’d personally like to build a feature which allows you to use simulation to produce scenarios as a starting point and then tweak them to your liking.
For MapCraft, I'm not up to speed on what it is now capable of. I am familiar with their cofounder’s previous work while at UC Berkeley. This is a good reminder for me to circle back on what MapCraft is up to - thanks! Also, if you’re a customer of theirs I’d love to hear how you think UrbanFootprint compares.
I'll have to leave questions about PECAS for someone else on my team as I don't personally have any knowledge/expertise on that one.
You mention population growth modeling and the linked article here mentioned the tool can’t predict changes in rents... is there any interest in incorporating some sort of lifecycle analysis or dynamic modeling like that here?
If that's what you're interested in, you might want to check out https://www.podaris.com/, which started as a tool for planning autonomous Personal Rapid Transit networks. It now supports conventional modes as well as theoretical systems (hyperloops, etc.)
(Disclaimer: I'm the founder. Not meaning to steal UrbanFootprint's thunder, as it looks like a great tool, and it would be great to interoperate with them someday.)
I believe the tool works, but I'm skeptical it will have any impact. There's a lot of "no shit sherlock" development that isn't happening. Building dense near the train station is literally textbook urban planning. It shouldn't take a complex simulation to prove it. The problem I see is that the electorate doesn't work the way the founders imagine.
>When communities can see comprehensive data about multiple plans for the future, the startup’s founders say, it becomes easier to compare them and reach consensus.
What actually happens is people like and dislike certain things and will post hoc justify entrenching their preferences in law.
You're not wrong inasmuch as that's what happens today. However: tools do help people think. I've worked as both an architect and an urban planner, and seen how people engage with complex systems in those situations. Absent good modeling tools, it's easy to spin fantasies and post-hoc justifications in virtually any direction, and when those directions diverge within a collaborative effort, nothing gets done. A good modeling tool can help the situation simply by limiting the degree of divergence that is possible.
It's a nudge, and it'll take the culture a long time (if ever) to fully respond to it, but it's definitely a step in the right direction.
Even given the dynamics you're describing, the basic idea still makes intuitive sense to me: it's easier to convince someone with a pretty & intuitive simulation than it is without.
Your case would require complete rejection of rational arguments by the population. And while I do enjoy indulging in gratuitous negativity from time to time, reality clearly isn't quite as bleak as one may think. As but one example: power generation & distribution are one of the segments of infrastructure that often lead to controversies, but supply has so far managed to remain stable.
The optimist might say we're using some slack in the system to take our time to find solutions that might actually better than first drafts. Yet even the pessimist will at least agree that there's still a lot of potential to get worse.
Or good chunk of the people complaining about US electorate in 2018. It's easier to blame everything on evil Russian hackers manipulating elections than to accept that for a lot of people, given their particular problems and issues, this particular candidate may have been the least worst one, or promised the best things.
A different way to phrase this might change the assumptions. Demand for funds to do many civics projects in California and elsewhere is legally bound to involve stakeholders before funds can be considered. This software environment builds detail-rich mapping, both current conditions and also a proposal (or three). The result is a quantitative data set that can be visualized and queried. That data environment is on the table of each stakeholder, so that discussion can be grounded with detail, and attention put to specific content.
The scenario-sketching is available for private developers, public actors and government too, for their respective processes. SO it is not the tool that makes change.
> The problem I see is that the electorate doesn't work the way the founders imagine.
Ding ding ding! Precisely. We’ve been designing cities well for thousands of years. I mean, pre-automobile, it made zero sense to have 5, 10, 15, 20 miles between you and a merchant or a doctor or law enforcement.
This is a political problem. NIMBY property owners are increasingly the only ones able to vote in a given location, partially because they’re the only ones who can afford to live there. I mean, Jane Jacobs was writing in the ‘60s about how we’re ruining thousands of years of urban planning knowledge.
Of course US urban planning, and especially its reliance on cars and the wrongheaded deprecation of mixed zoning are a problem.
But telling people "you're idiots" just doesn't seem to change any minds. It being a political problem (which is almost a truism) doesn't change anything regarding the tools that can be successful.
NIMBYs obviously fear change. If they are right, and some development would cause unreasonable harms, then it'd be wrong to dismiss them. If, however, as most people seem to assume, their fears tend to be overblown, then what better way to change minds than showing them a believable simulation? Possibly one that lets them interactively fiddle with the assumptions, examine the experts' opinions previously presented as inscrutable dogma, and arrive at more realistic expectations.
Was it even tried? I'm yet to see "believable simulations" shown to the general public. Something that doesn't look like a video of an indexed-color Earth with randomly rotating palette, but is explainable, explained, and possibly interactive.
I myself believe scientists working on this have such simulations. But I've never seen any being discussed with regular people.
Speaking only and solely for myself, I've seen a whole series of believable simulations shown to the general public and explained to a degree I find adequate. VESL springs to mind: https://vesl.jpl.nasa.gov/
Of course, believable tends to be a matter of opinion, and climate scientists have often been treated somewhat harshly by regular people when they attempt to engage.
It's enough to lead me to think that there is a whole series of such tools, already existing and meeting every one of your completely reasonable criteria! Unfortunately, they do not seem to be having the impact an optimist might hope for. Among other things, it requires a person to trust the programs involved, trust that the models correspond meaningfully to reality, and trust the intentions of the people creating them.
None of these are small asks in a politically charged arena, you know?
Haven't seen or heard that before; thanks, I'll have something to show some people.
> Among other things, it requires a person to trust the programs involved, trust that the models correspond meaningfully to reality, and trust the intentions of the people creating them.
Yes, those are the basics, they may not be small asks in a politically charged areas, but if you don't have them your alternative is to coerce people by threats. It's entirely reasonable to require these three things, as if people were to suddenly suspend them, they'd be quickly conned. This makes the problem hard, especially that you have a lot of people, including pretty much all news companies, who make money by burning that trust.
Maybe a lot of convincing simulations were shown as a matter of large-scale public discussion, and were rejected based on trust issue. If so, I must have missed it, and it's a shame :(.
RE politically charged areas, I'm reminded of this quote:
> As near as I can tell, the political process converts ideas from logical propositions into group signaling devices. Once a idea becomes a signal, any question of its truth or falsehood is commonly ignored for several human generations.
Cities Skylines with the Real Time Mod and Transport Manager President Edition mods added
Gets pretty close to simulating what cities go through (Foreign Affairs not withstanding) and makes a great tool in communicating to residents and businesses (as the Swedes and Finnish discovered).
A basic use case is a city updating its General Plan, which would start with a forecast of how much population growth is anticipated / needs to be accommodated. A planner then needs to assess where new residents will live, work, shop, and play. Perhaps even more essential, how are people going to travel between all of these activities? Will the new growth be auto-dependent, transit-focused, walkable? Is any of the existing or planned development in hazard areas such as flood of wildfire? What are the energy and water use impacts of the plans?
We’re using Python and Postgres/PostGIS on the backend to answer these questions and a React SPA to serve it up and make it interactive in a browser.
Also, if you happen to live in California, UrbanFootprint is available for free to your city through our California Civic Program (http://info.urbanfootprint.com/california-civic-program) so feel free to nudge them to get in touch ;)