CTPP May 2023

ctpp@listserv.transportation.org

3 participants
4 discussions

FW: Datapalooza! Call for content! Please distribute!

by Penelope Weinberger

Hello CTPP Fam, This is a little outside our regular types of emails, but perhaps you or someone in your organization has something to share. Are you working on a data project that's neat, cool, or keen; one that has you excited and makes you glow with pride? Perhaps someone in your DOT or organization is. Showcase this project at the AASHTO Committee on Data Management and Analytics Datapalooza! Datapalooza is a two day remote/virtual event August 22 and 23, 2023. We are seeking Innovative, Collaborative, Educational (ICE) content to cool us off in the heat of August. Send us a description of your: * Demonstration Projects * Analytics Projects * Rapid Innovation Projects * Any other super interesting projects! Please use the link below to submit an abstract by June 2nd https://www.surveymonkey.com/r/DatapaloozaContent Penelope Weinberger Program Manager for Transportation Data, AASHTO Data.transportation.org

2 years, 2 months

Obtaining Census Population Estimates using the R package tidycensus

by Charles Purvis

I’m forwarding a “toot” (Mastodon equivalent of a tweet) regarding an updated version of the r package tidycensus. All of the hub bub in the current news is about the new vintage 2022 population estimates released by the Census Bureau this past Thursday, May 18th. Lots of interesting stories are being told!! I tested tidycensus for California and US states, counties, and places, and it works great. It’s a little awkward since you need to split your analysis into the 2010-2019 period from the 2020-2022 period (i.e., can’t get 2019 thru 2022 in one call!) My tidycensus population estimates r script is available here: https://github.com/chuckpurvis/r_scripts/blob/main/popestimates_calif_01.R ########### UPCOMING: The “SF1” equivalent to Census 2020 is embargoed right now (media has access to the data; public, doesn’t) and will be released this Thursday, May 25th. This is the “DHC” "Demographic and Housing Characteristics” file. My expectation is that Professor Walker will have updated tidycensus to read DHC by the end of my morning coffee on Thursday. ########### “A brand-new version of the #rstats tidycensus package is now on CRAN, supporting the brand-new 2022 Population Estimates (which you can't get from the API). Download the new version today and start making charts like this!” - author Kyle Walker https://walker-data.com/tidycensus

2 years, 3 months

Comparison maps for San Francisco City: Population Density (PL94171) and Job Density(LEHD)

by Charles Purvis

Attached are two maps created using the R package ggplot2. The data is from Census 2020, PL 94-171 for population density (persons per square mile) and the Census Bureau’s LEHD/LODES estimates for 2020 (total jobs per square mile). My six “density categories” are defined as follows: dplyr::mutate(density_grp = case_when( jobs_per_sqmi < 500 ~ "1.rural", jobs_per_sqmi < 1000 ~ "2.rural-suburb", jobs_per_sqmi < 6000 ~ "3.suburb-disperse", jobs_per_sqmi < 10000 ~ "4.suburb-dense", jobs_per_sqmi < 20000 ~ "5.urban", jobs_per_sqmi >= 20000 ~ "6.urban core", TRUE ~ "0.missing data”)) “Mutate” sounds horrible, but it’s just the “dplyr” verb for “create a new variable”. (“dplyr” is a very handy-dandy R package). Certain block groups in the City of San Francisco have “rural” population density. These are our parks: Golden Gate Park, Lincoln Park, McLaren Park. The Presidio might be categorized as a “dispersed suburb”. Most block groups in San Francisco are “urban core” (greater than 20,000 persons per square mile). Some block groups in downtown San Francisco are not that heavily populated, such as the Financial District and Union Square areas. (Of course, vast overwhelming anecdotal evidence suggests that downtown San Francisco is a ghost town due to the pandemic. Ain’t necessarily true, of course.) But the San Francisco financial district could not be considered a “suburb” by any stretch of the imagination. This is why we should also review data on jobs per square mile (job density) as a counter measure to simple population density. My resulting hybrid measure would use the maximum value of population density (total population per square mile) and job density (jobs per square mile) in determining if a neighborhood is rural, suburban, or urban. I would use the same density ranges as shown above. I could use a compound density measure (e.g., (population + jobs) / land area), but I’m satisfied with using the maximum of population OR jobs to typify a neighborhood. (Our San Francisco financial district block group - tract 117.00, block group #3 - bounded by Bush-Kearny-Sacramento-Drumm-Market, is 0.07984 square miles, with a population of 442 and a job base of 49,634 employees, for a population density of 5,535 persons per square mile and a job density of 621,613 jobs per square mile. Yes, over 600,000 jobs per square mile!!) Hope this is of interest. Chuck Purvis, Hayward, California

2 years, 3 months

Updated R package lehdr released May 14th, for LODES/LEHD analysis

by Charles Purvis

The R package “lehdr” created by Professor Jamaal Green at Penn (and others) was updated on 5/14/2023 to include year 2020 data from the Census Bureau. Yay!! This is a great way to acquire (free) data on workers-at-block (block group, tract, county, state) of residence and of workplace. The user can then link the data with census GIS files using the R program “tigris” and map and display all sorts of data. The LEHD/LODES data also has origin-destination “flows” (block-to-block, block group-to-block group, tract-to-tract, county-to-county, state-to-state). My r script using lehdr and tigris for California is free to share, here: https://github.com/chuckpurvis/r_scripts/blob/main/lehdr_California_2020.R My r script for using tidycensus to map population density (Census 2020, PL 94-171) for the Bay Area and San Francisco City is here: https://github.com/chuckpurvis/r_scripts/blob/main/PL94171_Calif_PopDensity… The “lehdr” ( pronounced “lee-ter”) package is used to download and format data from the Census Bureau’s LEHD / LODES program. ## LED = Local Employment Dynamics ## LEHD = Longitudinal Employer Household Dynamics Program ## LODES = LEHD Origin-Destination Employment Statistics ## UI = Unemployment Insurance ## QCEW = Quarterly Census of Employment and Wages ## RAC = Residence Area Characteristics ## WAC = Workplace Area Characteristics ## CES = Center for Economic Studies, US Bureau of the Census Documentation on LEHD/LODES is available from the Census Bureau here: https://lehd.ces.census.gov/data/#lodes https://lehd.ces.census.gov/data/lodes/LODES8/LODESTechDoc8.0.pdf Also of interest to MPOs and State DOTs will be the NCHRP report published 12 years ago (September 2011): https://onlinepubs.trb.org/onlinepubs/nchrp/docs/NCHRP08-36(98)_FR.pdf “Improving Employment Data for Transportation Planning” Hope this is of interest! Chuck Purvis Hayward, California ###

2 years, 3 months

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

CTPP May 2023