Geo Spatial Analysis

[shapironick]

I made a quick outline of spatial data analysis that could be performed now (i.e. before manual coding is completed) using HIFLD. While the ECHO data provides valuable information about the hazards produced by the prisons, we also need to assess exposures coming from proximate sources of exposure.

In terms of geography, we're interested in understanding two things: 1) are prisons, on a national level, more likely to be located in proximity to hazardous sites? This analysis would work towards a journal article on the state of toxic prisons. 2) which prisons specifically are close to toxic infrastructure? so that we can: a) cross check them with their status in the ECHO database to potentially demonstrate the limitations of using ECHO alone to assess exposure b) to create a list of prisons where we might want to do in-situ environmental monitoring to ground-truth exposure and to potentially reach out to incarcerated people in those prisons.

The variables to compare the prison location data with are (I wonder how much of this would be easier accessed through EJ screen): 1) Airport locations here (this will help us understand three exposures 1) lead from avgas 2) PFOS from AFFF firefighting foam 3) noise (which relates to both mental and cardiac health). We could filter for only airports in the Fac_Type column. 2) Military facilities. I had some trouble finding this data. But here is the best I could do. Its an archived version of military data. You can scroll down to the bottom where it says "Downolad [sic] Geospatial Information for U.S. Military Installations, Ranges, and Training Areas" zipped shapefile 6.5 MB) or just click here. 3) Brownfield locations. Which can be found here. 4) Superfund locations. Which can be found here. 5) There is no authoritative map of PFAS contamination but colleagues at northeastern have put together this list. But its not geocoded

[klfranco]

Update 2/18/2020. We worked on question 1 in our work session tonight. In our attempt to map carceral facilities in relation to Brownfields and Superfund sites we ran into data access issues. After troubleshooting, we finally got the appropriate data to start in on this sub-project. Next week we will work on producing a preliminary map that has prisons mapped alongside with these toxic sites.

[shapironick]

🙌🙌🙌🙌 Amazing!! Thank you Konrad and team!

Sent on the move

On Feb 18, 2020, at 7:54 PM, Konrad Franco notifications@github.com wrote:

 Update 2/18/2020. We worked on question 1 in our work session tonight. In our attempt to map carceral facilities in relation to Brownfields and Superfund sites we ran into data access issues. After troubleshooting, we finally got the appropriate data to start in on this sub-project. Next week we will work on producing a preliminary map that has prisons mapped alongside with these toxic sites.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[shapironick]

A small note: I was talking w/ one of the geospatial research staff at the library here and she said that UCLA has access to ESRI ArcGIS geocoding service. if that can be helpful for the PFAS data just let me know and she can connect me to the campus rep.

Big thanks y'all!

[lindsaypoirier]

What would be considered "in proximity"?

[shapironick]

this article might be a good starting place https://ehp.niehs.nih.gov/doi/pdf/10.1289/ehp.02110s2183

it looks like the number of BFs per census track and also acreage are important analytical frames. So this might be a bit beyond just straight proximity. (in one study I saw low birth weights were associated with concentrations of BFs but not necessarily proximity)

for superfund sites I keep seeing 4 miles.

I think for our spatial analysis we will want to think about total exposures and not break down each potential exposure pathway, as that's not how they hit the flesh. the question to me is: how do we assess the exposome in total via data?

Does it make sense to start aggregating multiple variables first and conducting some proximity and census concentration/acreage analysis after we've built up all the exposure routes more?

On Wed, Mar 11, 2020 at 11:10 AM Lindsay Poirier notifications@github.com wrote:

What would be considered "in proximity"?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Carceral-Ecologies/Carceral-ECHO-data/issues/17#issuecomment-597786288, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABZH332HXJ66IG7TI65HZOTRG7HVFANCNFSM4KWWMXVA .

-- Nicholas Shapiro Assistant Professor UCLA Institute for Society and Genetics Office: (310) 206-2366

[shapironick]

ported this thread over to the correct (new) repo