ACS Alchemist: Easy Access to Census Data

ACS Alchemist: Easy Access to Census Data

A year ago we started building data collection tools to support predictive policing research in conjunction with Temple University’s Center for Security and Crime Science.  The NIJ-funded research is aimed at combining socio-economic data that changes over long period of time with short-term forecasts.  The result will be a new technique for forecasting crime risk.

For the first phase of the research, we need to generate socio-economic data that is accurate, current and can be localized to the neighborhood level.  The project team has turned to the US Census Bureau American Community Survey (ACS).  The ACS  is an ongoing statistical survey that provides data in the years between the decennial census (it replaces what used to be known as the “long-form”).  The ACS is the basis of a broad range of research, and it invaluable source to get a general picture of everything from the economy to family structure.

The ACS is very useful, but it’s somewhat complex to extract the data.  That’s why we’re excited to announce the release of ACS Alchemist as a free, open source project.  ACS Alchemist is a user-friendly tool that enables you to extract data from the ACS directly into a Shapefile for display on a map or for further analysis.  We’re finding the tool useful internally because it makes it extremely easy to access Census tract and block group level estimates in a format readily usable in many GIS packages.

Our research with Temple University is focused on crime, but what other questions can the ACS help answer?  There are thousands of variables available , but here are a few that we found interesting.   (All estimates below are from the 2010 ACS 5-year estimate aggregated to the county level)

What proportion of people walk, bike or take public transit to work? 

  • Philadelphia County –  35.9% [walk: 7.1%; bike: 1.4%; public trans: 27.3%]
  • Allegheny county (Pittsburgh) – 13.5%  [walk: 4%; bike: 0.3%; public trans:9.1%]
  • Jefferson County – 4.2% [walk: 3.7%; bike: 0.2%; public trans: 0.4%]

Of the people that walk to work, what percentage own at least one car?
Percentage estimate of those walking to work who own 1 car or more

  • Philadelphia County –  59%
  • Allegheny county (Pittsburgh) – 72%
  • Jefferson County – 85%

Of the people that walked to work, how long did they walk?
Percentage estimate of those walking to work who traveled between 0-15 minutes

  • Philadelphia County – 50% walked less than 15 minutes
  • Allegheny County (Pittsburgh) – 63% walked less than 15 minutes
  • Jefferson County – 89% walked less than 15 minutes

During the next phase of this project we’ll be building on the ACS Alchemist to model year-long crime levels based on changing neighborhood demographics.   Stay Tuned!  In the meantime, if you want to use neighborhood-level Census data, grab a copy of ACS Alchemist from GitHub today!


This project was supported by Award No. 2010-DE-BX-K004, awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. The opinions, findings, and conclusions or recommendations expressed in this software are those of the author(s) and do not necessarily reflect those of the Department of Justice or Temple University.

The software was developed by Azavea in connection with the National Institute of Justice grant awarded to Jerry Ratcliffe and Ralph Taylor of Temple University’s Center for Security and Crime Science.