eli knaap
LightNews LightNews
banner
knaaptime.com
eli knaap
@knaaptime.com
580 followers 1.1K following 220 posts
urban planning & public policy prof @ UC Irvine | core dev @ PySAL & QuantEcon | urban social science & spatial data science | open source

https://knaaptime.com
Posts Replies Media Videos
knaaptime.com
instead i get

😑
<folium.folium.Map at 0x34e1cff80>
November 19, 2025 at 5:37 PM
knaaptime.com
weighted results are also simple (70k observations this time). The histogram is still different from the original graph though (and the median rent actually drops a bit to 2270)

@besttrousers.bsky.social are you doing something i'm missing?

bsky.app/profile/knaa...
occ_rent = occ_rent.set_index('CONTROL') weighted_idx = occ_rent.index.repeat(occ_rent['FW']) weighted_idx weighted_rent = occ_rent.merge(weighted_idx.to_frame(), left_index=True, right_index=True, how='right') weighted_rent.RENT_AMOUNT.hist(bins=20) weighted_rent.RENT_AMOUNT.describe()
October 7, 2025 at 8:28 PM
knaaptime.com
extremely quick and dirty, but ignoring the survey weights, you can basically recreate these numbers from NYCHVS in ~9 lines. Only gives you 138 observations and the histogram is a little different (original seems to include $0? among a few other diffs), but the 2300 number shows up, fwiw.
import pandas as pd units = pd.read_csv("https://www.nyc.gov/assets/hpd/data/allunits_puf_23.csv") # brooklyn 3br units = units[(units.BORO==2) & (units.BEDROOMS==3)] # "market rate" units = units[units.CSR==80] # occupied table has rents occ = pd.read_csv("https://www.nyc.gov/assets/hpd/data/occupied_puf_23.csv") occ = occ.set_index('CONTROL') occ_rent = occ.merge(units[['CONTROL']], left_index=True, right_on='CONTROL') occ_rent = occ_rent[~occ_rent.RENT_AMOUNT.isin([0,-1,-2])] occ_rent.RENT_AMOUNT.hist(bins=20) occ_rent.RENT_AMOUNT.describe()
October 7, 2025 at 7:42 PM
knaaptime.com
one thing i do miss from twitter is having people like Gillies around (he's not here, right?)

also: it's bullshit.
post from sean gillies blog https://sgillies.net/ I've been on the Project Steering Committee (PSC) of the GDAL project for a few years now. It's a hugely significant software project, one that I'm very fond of, and I like to think that I've had an impact in my time on the committee. Recently, though, the steering part has become less rewarding. By granting Amazon S3 and its imitators special, featured storage protocols, GDAL is stuck carrying water for big and lucrative cloud service providers and a few hyper-scaling customers that don't adequately support the project. It's a crappy situation, not unique to GDAL. A lot of open source projects find themselves ruthlessly exploited these days. This is my personal view. I don't speak for GDAL. More importantly, I'm doing less geospatial work these days, and thus I'm increasingly uncertain about the rapidly-changing currents in which the PSC is steering the project. Do things like "GeoParquet" and "GeoArrow" really matter to practitioners who aren't working for a handful of hyper-scaling organizations? Is "GeoAI" for real or is it bullshit? Do we steer the project towards these interests or away from them? I truly don't know. I wish I did. Can I fully engage with the problems that GDAL faces over the next year? That's my criteria for participating in the PSC. I can't, and so it's time to go. The project is in good shape, I believe, and I hope that I'm making room for someone with a great vision to join it.
October 3, 2025 at 9:04 PM
knaaptime.com
arial photo of armitage ave from google maps; rico fresh is adjacent to funston elementary school
October 3, 2025 at 6:15 PM
knaaptime.com
incidentally, about 3 blocks (in the other direction) from where this was taken back in 2017 knaaptime.com/urban_analys...
brown boston terrier standing on a side walk with "fuck trump" etched into the concrete
October 3, 2025 at 5:57 PM
knaaptime.com
Among so many other examples, this is just not how governing works
The White House took steps to maximize the pain of the government shutdown, halting funds for democratic-led states and readying mass layoffs
October 2, 2025 at 12:51 AM
knaaptime.com
cool news from PySAL world: the (imminent) next release of the `esda` module will have (1) spatial correlograms and (2) (more importantly) @levijohnwolf.bsky.social's partial Moran's I estimator described in

www.tandfonline.com/doi/pdf/10.1...

(both already in main if you can't wait 😎)
relationship diagram showing different spatial interaction structure(s) underlying alternative concepts of spatial correlation and the Moran's I coefficient, as shown in https://www.tandfonline.com/doi/pdf/10.1080/24694452.2024.2326541
September 25, 2025 at 6:35 PM
knaaptime.com
i mean, he's also quite clear about *why* he's using a sensational title in his into. He wants this piece to be part of the conversation. But if you think clickbait is too pejorative, i guess i undertand
Given the title, there’s little doubt that the provocation was intended; after all, a piece entitled “Why America should carefully extend urban infrastructure to accommodate growth” might work for the Congress for the New Urbanism, but wouldn’t generate much buzz.
September 11, 2025 at 10:30 PM
knaaptime.com
i dunno. I thought dougherty's piece was wildly off, but this reads as a little disengenuous.

compact sprawl? by definition that's *not sprawl*. The cogent argument is we need metropolitan-scale regulation to support infill. It's better to just say that than to be click baity for more reads
The suburban growth in the Sunbelt may be “sprawl,” if one is going to use that term to describe any and all development outside already built-up cities, but as Figure 5 from Princeton, Texas (a city featured in Daugherty’s article) shows, it is very compact sprawl. The homes shown in the image are built at a gross density including streets, playgrounds and such, of roughly five units per acre, and a net density in each block with 6250 square foot lots, of seven units per acre. Not quite downtown apartment densities, to be sure, but if one accepts that there is a continued demand for single family homes, a highly efficient use of land, and comparable to the densities of older single family neighborhoods in cities like Cleveland.
September 11, 2025 at 9:52 PM
knaaptime.com
not yet published but I'm talking about our 3rd paper on the project at WVU's Regional Research Institute next friday, describing how to combine discrete choice with location-allocation models to explore tradeoffs in siting harm reduction services. Happy to post slides if anyone interested
a map of riverside county CA showing four potential harm reduction sites and their service areas, assuming a 5 kilometer service area a map of riverside county CA showing four potential harm reduction sites and their service areas, assuming a 15 kilometer service area
September 8, 2025 at 7:39 PM
knaaptime.com
hey, um. esri is not a valid source for population forecasting.

www.nytimes.com/interactive/...
Those gains will come at the expense of major Democratic states like New York and California, according to a New York Times analysis of population projections by Esri, a nonpartisan company whose mapping software and demographic data are widely used by businesses and governments across the world.
August 26, 2025 at 4:31 AM
knaaptime.com
OR, i'll forfeit the 20mil if you promise to stop being a dick to everyone else fighting for justice. This dude is incredible

bsky.app/profile/the-...
Columbia University graduate Mahmoud Khalil is seeking $20 million in damages from the Trump administration, according to court records — or an apology and rescission of the government’s policy of targeting pro-Palestine student demonstrators. (Reuters)
July 10, 2025 at 9:37 PM
knaaptime.com
continues to cackle from sn diego

bsky.app/profile/opin...
June 20, 2025 at 6:42 PM
knaaptime.com
we're just citing blog posts in SC rulings now? ill never understand this in the same way ill never understand why people are willing to read yglasias. who the fuck cares what ja rule thinks at a time like this?

bsky.app/profile/ripp...
June 18, 2025 at 11:36 PM
knaaptime.com
this is the distinction between people who understand AI and people who don't.

if you think the computer is capable of PhD caliber work (let alone 12x PhD work) you understand neither LLMs nor doctoral degrees (or people who hold them)
ROOSE The mental model I sometimes have of these chatbots is as a very smart assistant who has a dozen Ph.D.s but is also high on ketamine like 30 percent of the time. But also, the bar of 100 percent reliability is not the right one to aim for here: The base rate that we should be comparing with is not complete factuality but the comparable smart human given the same task.
June 18, 2025 at 1:59 AM
knaaptime.com
Galster calls this influence from multiple externality spaces; we call it 'multiple spatial embeddings'. Travel behavior matters for the definitions of some externality spaces, but residential location defines your participation in many of the most important ones

link.springer.com/article/10.1...
image from George Galster's *Making our neighborhoods, making our selves" showing partially overlapping polygons representing multiple conceptual 'externality spaces' for two different people
June 16, 2025 at 2:36 AM
knaaptime.com
oh, the other reason is neighborhoods really only differentiate along a few dimensions, which are essentially class, race, age/household size, and density

knaaptime.com/urban_analys...

bsky.app/profile/knaa...
"factorial ecology maps of the chicago metro region, showing maps of dimensions "SES", "segregation, "family structure" and "urbanization" "factorial ecology maps of the LA metro region, showing maps of dimensions "SES", "segregation, "family structure" and "urbanization"
June 11, 2025 at 5:20 PM
knaaptime.com
When you’re a “conservative” being chastised by faculty at the Scalia School 😳

www.nytimes.com/2025/06/10/u...
“He is declaring utterly bogus emergencies for the sake of trying to expand his power, undermine the Constitution and destroy civil liberties,” said Ilya Somin, a libertarian professor at Antonin Scalia Law School who represents a wine importer and other businesses challenging some of Mr. Trump’s tariffs.
June 10, 2025 at 7:10 PM
knaaptime.com
summer 25 addendum 🙄

bsky.app/profile/knaa...
As the writing of this book closes and the research infrastructure in the U.S. crumbles, it is important to clarify that part of my work on this collection of materials was funded by NSF POSE Grant #2345820, a project designed to make spatial data science more transparent and accessible for a broader range of users and contributors. This work is made open and freely available, explicitly, so that others can build on top of it to the scientific benefit of us all. It would not be possible without public research funding. More importantly, the vast majority of software packages demonstrated here have been supported directly and indirectly by a string of more than eight NSF grants, as well other funders including the Bill and Melinda Gates Foundation, the National Institutes for Justice, and the National Institute for Health3. geosnap was conceived during my postdoc on spatial distribution dynamics for neighborhoods (NSF #1421935 and #1733705), and tobler, segregation, and spopt were all borne from our RIDIR grant (NSF #1831615) at UCR–and that’s just since I’ve been on the team! Multiple team members before me also received NSF dissertation grants. Today the PySAL stack is cited in everything from genomics (Moses et al., 2023) to astrophysics (Martel et al., 2024)—classic agglomeration from knowledge-spillover, generating a massive return on the original investment—and that only happens because of public funding.
June 10, 2025 at 4:24 PM
knaaptime.com
the culprit here is variation in the size of geographic units being displayed. One possible way to try to overcome that is to interpolate to a consistent zonal system

bsky.app/profile/paul...

knaaptime.com/urban_analys...
choropleth maps of DC using the same population data; the right hand side shows original units, the left hand shows a regular hexagonal grid
June 9, 2025 at 10:35 PM
knaaptime.com
there's a ~complete rough draft of Urban Analysis up on my website (as in, code executes all the way through)

knaaptime.com/urban_analysis

still lots of editing to do, but if you're into cities, Python, or spatial analysis, give it a look and let me know what's wrong :)
table of contents for Urban Analysis and Spatial Science Introduction Spatial Data 1 Data Sets 2 Geoprocessing 3 Geovisualization Interpolation and Change of Support 4 Areal Interpolation 5 Dasymetric Interpolation 6 Boundary Harmonization Spatial Analysis 7 The Spatial Graph 8 Spatial Autocorrelation 9 Space-Time Event Clustering Accessibility and Interaction 10 Travel Isochrones 11 Accessibility 12 Multimodal and More Urban Spatial Structure 13 Location and Land Use 14 Concentration and Agglomeration 15 Employment Centers Segregation 16 Segregation Measures 17 Space, Scale & Dynamics 18 Single and Comparative Inference Urban Factor Models 19 Latent Spatial Processes 20 Factor Ecology 21 Ecometrics and Composite Indices Neighborhood Delineation 22 Social Areas & Urban Typology 23 Geodemographic Clustering 24 Regionalization Neighborhood Change 25 Spatial Dynamics in Math Achievement 26 Stochastic Cellular Automata 27 Neighborhood Dynamics Spatial Econometrics 28 Applied Spatial Econometrics 29 The Hedonic Housing Price Model 30 Hedonic Models in Space Location Models 31 Location Models 32 Location Choice 33 Spatial Optimization graph of economic "bid-rent" model showing the relationship between land rent and distance from the market center for three types of land consumers
June 9, 2025 at 6:10 PM
knaaptime.com
dear political scientists,

who are your people that inherit from the Hotelling tradition and what do they have to say about today? why are democrats the only ones still playing the hot dog game???

seriously

bsky.app/profile/jame...
excerpt from "Stability in Competition" (Hotelling 1929): So general is this tendency that it appears in the most diverse fields of competitive activity, even quite apart from what is called economic life. In politics it is strikingly exemplified. The competition for votes between the Republican and Democratic parties does not lead to a clear drawing of issues, an adoption of two strongly contrasted positions between which the voter may choose. Instead, each party strives.to make its platform as much like the other's as possible. Any radical departure would lose many votes, even though it might lead to stronger commendation of the party by some who would vote for it anyhow. Each candidate " pussyfoots," replies ambiguously to questions, refuses to take a definite stand in any controversy for fear of losing votes. Real differences, if they ever exist, fade gradually with time though the issues may be as important as ever. The Democratic party, once opposed to protective tariffs, moves gradually to a position almost, but not quite, identical with that of the Republicans. It need have no fear of fanatical free-traders, since they will still prefer it to the Republican party, and its advocacy of a continued high tariff will bring it the money and votes of some intermediate groups.
June 7, 2025 at 6:23 PM
knaaptime.com
without question every UC and CSU is an agglomeration center... i'd expect the local economic hit on places like Riverside or Merced would be massive; the hit + *spillover* from the LA, SF, SD metros would be massive

www.cnn.com/2025/06/06/p...

bsky.app/profile/knaa...
employment centers in the San Diego region with SDSU and UCSD campuses circled
June 6, 2025 at 8:56 PM
knaaptime.com
HUD is iconic. favorite building ive ever worked in
the Robert Weaver bulding, home of the Department of Housing and Urban Development
May 29, 2025 at 4:27 PM