The most recent free street data that I have found is the StreetMap data (based on 2000 Tigerline data) that comes with an ArcView license. As you can imagine, this data is particularly ill-suited for geocoding addresses in areas that have experienced substantial new development in the last eight years. Luckily, I just found a great web application that allow you to obtain X/Y (latitude/longitude) coordinates based on street address. Better yet, it allows you to batch process the addresses that you want to geocode.

The REP is currently geocoding the addresses of bank foreclosures in Sacramento County. Using StreetMap data (based on 2000 Tigerline data) I was only able to geocode 20% of the 4,681 addresses that I had. Using batchgecode.com, I was able to gecode 4,674 of the 4,484. Not bad for a free web app considering that newer street data can easily cost you over $8,000.

We made a few new additions to our resources today that we think you should take a look at:

The Kirwan Institute for the Study of Race and Ethnicity

• Site Summary: The Kirwan Institute for the Study of Race and Ethnicity’s “ultimate goal is to stimulate and facilitate transformative change to bring about a society that is fair and just for all people.” It approaches this goal “with the understanding that all communities of people are interconnected and that by fostering positive relationships through research, scholarship and collaboration, we will deepen the possibility for change.” The Kirwan Institute is a true pioneer in the fight to achieve equitable results for communities of color and its research and advocacy, particularly around opportunity mapping, represents some of the most innovative and effective advocacy currently being undertaken by advocates for race equity. The Institute also posts all of their PowerPoint presentations on their site. Available presentations cover topics ranging from GIS to framing to structural racism.

HealthyCity

• Site Summary: HealthyCity offers perhaps the most comprehensive access to community resources, demographic/health data, and cutting edge online GIS mapping technology that the REP has ever seen. For the time being, the site only offers geographic coverage for Los Angeles county. We hope that HealthyCity will be going statewide soon but until that time we will have to stew in our jealousy of the wonderful online mapping and data analysis tools that residents of the city of angels have access to.

• Suggested Uses: If you have any mapping or data analysis needs related to Los Angeles county and you are not adverse to free, powerful, user-friendly online mapping and data analysis tools than HealthyCity is for you.

We have spent much time exploring the many ways in which advocates can use mapping to highlight the social isolation and segregation from opportunity of their client communities, among other things, and advocate for positive change. (See REP Blog mapping archives.) However, mapping is tool available to everyone and, as this story may suggest to you, can have its downside, especially when applied to groups racialized as prone to radicalism and violence.

According to the NY Times in Protest Greets Police Plan Map Muslim Angelinos, the Los Angeles Police Department (LAPD) plans to map the location of southern California Muslims under the auspices of the “Violent Radicalization and Homegrown Terrorism” Act. The “Violent Radicalization and Homegrown Terrorism” Act defines “ideologically based violence” as “the use, planned use, or threatened use of force or violence by a group or individual to promote the group or individual’s political, religious, or social beliefs.” That’s the place from which LAPD’s assumptions about Muslims run wantonly in the direction of implicit bias. As many are aware, the law of implicit bias is unsettled.

Though not listed among the LAPD’s official maps, the department never officially withdrew its proposal to map SoCal Muslims. Here’s what others had to say about the proposal.

Peter Bibring, a lawyer with the A.C.L.U. of Southern California, expressed the alarm many felt at hearing that the “starting point for a police investigation is ‘let’s look at all Muslims.’”

Mike German, policy counsel for the ACLU, called the plan “wrongheaded” because the bill calls for heightened scrutiny of people who believe, or might come to believe, in a violent ideology, which, the In These Times article notes, is perfectly legal.

Hussam Ayloush, executive director of the Greater Los Angeles Area chapter of the Council on American-Islamic Relations (CAIR-LA), debates an LAPD proposal to “map” Southern California Muslim communities on “NBC Nightly News.” See the video interview here. The video questions whether racial profiling can be good policing. For an historical view of the process of official state mapping of minority communities and racial profiling policy, see the LA Times article, “Community Profiling’s Long, Sad History” by Professor Richard Marcus of Cal State Long Beach.

Pr. Marcus points out that, under the Bush Administration:

The U.S. Department of Justice banned racial profiling, calling it unconstitutional. Under this definition, former Atty. Gen. John Ashcroft followed in February 2002, saying that using race “as a proxy for potential criminal behavior is unconstitutional, and it undermines law enforcement by undermining the confidence that people can have in law enforcement.” I guess the LAPD missed the memo.

The REP recently had the opportunity to use EPA data in a mapping project for CORE. We wanted to share with you two nifty datasets that we found.

The first data set is the CERCLIS Superfund database. This database contains information on all NPL and non-NPL superfund sites in the country. Don’t bother trying to geocode the non-NPL sites as the street data is horrible. The REP suggests using the latitude and longitude data to locate the sites in the CERCLIS database. If the Superfund data leaves you a little confused, take advantage of the marvelous Superfund Info.Line (703-284-8214).

The second data set is the 2006 Toxic Release Inventory (TRI). This database includes every facility in the country that stores or releases toxic chemicals. Be aware that the six different dbf. files that compose the database can be a bit tricky to link as a single facility can have multiple entries. However, almost all the information that you need is in the first dbf. file. The REP recommends locating the facilities using latitude and longitude as the street data element is less than perfect.

Good news on the 2010 Census front. The Bureau just published the final criteria for 2010 block groups in the Federal Register. Despite early rumors that the minimum population for a  block group would be raised, the Bureau decided to keep it at the 2000 level of 600 persons.

Moose, why are you sharing this inane information with us? Let me explain. First, data and mapping folks (myself included) need to know what to expect from the 2010 Census so that we can continue expose inequities. Second, an increase in the minimum population of a block group would have the effect of hiding small and insular communities (the communities we serve) so the news means that the Census will remain a uniquely powerful and complete data source for poverty advocates. Lastly, I figured we could all use some good news to mollify our sadness about the demise of the “long form.”

The demise of the census long form is common knowledge amongst Bureau fanboys and girls. However, we all could use a little more information about the big switch. For instance, did you know that the sample rate will change from 1 in 6 to 1 in 40? The current issue of ArcUser has a niffty article comparing the long form and the new ACS.

I also liked the article “Data Evaluation Tips: Buliding more accurate demographic databases.” Who isn’t interested in the pycnophulactic and dasymetric methods for converting data between different geographic levels? These methods’ scary names belie their practical applications in mapping poverty, race and opportunity.

LSNC and other Legal Services corporations have been inundated with foreclosure related issues in recent months. As most poverty attorneys who have handled a foreclosure case will attest, prevention and early intervention are key to preventing foreclosure. Of course, before you can deliver education and community outreach to LSC eligible homeowners you need to know were they live. You also need to identify where populations of LSC eligible homeowners of color live. In many areas, due to the greater frequency of asset accumulation of Whites, if you do not include a separate analysis of race and ethnicity you will end up serving mostly White LSC eligible homeowners and excluding homeowners of color.

Lets conduct this analysis for Los Angeles County, California, using free online mapping software and free census data extraction tools so that any Legal Services program can replicate this process for their service area if it has a computer, Microsoft Office, and an internet connection. The online mapping software we will use is DataPlace and the free census extraction tool is GIStools. For purposes of space, this tutorial assumes that you have created an account with DataPlace and have downloaded the software and relevant datasets for GIStools.

First, we need to identify what Census data to use. Since we want to identify areas where foreclosure services outreach will have the greatest impact, we need to find the areas in Los Angeles county where there is a high percentage of homeownership and where there is a high percentage of LSC eligible clients. Since we also want to make sure that our outreach effectively reaches all LSC eligible homeowners we also want to find areas where there is a high percentage of homeowners of color and a high percentage of LSC eligible persons.

My suggestion is to use three datasets from the 2000 Census, SF3, to identify areas for possible outreach - P088 (Population by ratio of income to poverty), H007 (Tenure), and H013 (Tenure, White alone not Hispanic). Since we will be using DataPlace to map the data, we need to obtain the data at the Census Tract summary level (the smallest unit of Census geography that DataPlace can map). If I was going to map the data using ArcGIS, I would probably want to obtain the data at the Block Group summary level (the smallest unit of Census geography SF3 data is available at). Using GIStools extract the P088 and H013 data for California at the Census Tract summary level. Open the dbf. files (one for P088, one for H007, and one for H013) using Excel and copy and paste the all the data sets into one Excel spreadsheet (view example). Delete all rows for Census tracts not in Los Angeles county. You can tell which rows are for tracts in Los Angeles County by looking at the FIPSCO column. The first two numbers in the entry (”06″) represent the state code for California. The next three numbers are the FIPS codes for the various coutys in California. Los Angeles county’s FIPS code is “037.” Simply delete all rows where the value of the FIPSCO entry is not “06037″ (view example).

Second, we need to combine some of the data fields to simplify our spread sheet. Columns P088001 - 10 represent the various data fields of P088. Take a look at the Cenus Bureau’s technical documentation of the 2000 Census SF3 file to see what each of these columns represent. Since we want to know the count of person’s below 200% of the poverty level (our approximation of LSC client eligibility), we will create a new column (LSCPOP) that is the sum of P088002 - 09. The values of LSCPOP give us the count of potentially LSC services eligible persons for every Census tract in Los Angeles county (view example). Since it would also be helpful to know what percentage of the population is LSC eligible, lets create a new column (PERLSC) and use Excel to divide LSCPOP by P008001 giving use the percentage of the population that is LSC eligible for every Census tract in Los Angeles county (view example).

Columns H007001 -03 and H013001-03 represent the various data fields of H007 and H013 respectively. Take a look at the Cenus Bureau’s technical documentation of the 2000 Census SF3 file to see what each of these columns represent. We are going to use these columns to calculate the count of homeowners, the count of homeowners of color, and various percentages related to these categories.

The count of homeowners in a Census tract is simply H007002. We need to create a new column (PEROWN) to calculate the percentage of the population in each Census tract that is a homeowner. The values for this column are obtained by dividing H007002 by H007001 (view example).

The count of White homeowners is simply H013002. We need to calculate a new column (PERWHITE) to calculate the percentage of homeowners who are White in a Census tract. The values for this column are obtained by dividing H013002 by H007002 (view example).

We need to create a new column (HOCOLOR) to calculate the count of homeowners of color in a given Census tract. The values for this column are obtained by subtracting H013002 from H007002 (view example). We also need a new column (PERHOCO) to calculate the percentage of homeowners who are homeowners of color in a Census tract. The values for this column are obtained by dividing HOCOLOR by H007002 (view example).

Before we move on to step three, lets “clean up” our spread sheet by deleting columns that we no longer need. Before we do this, copy the whole spreadsheet and “Paste Spacial” it into a new spreadsheet. Paste “Values” only. Take a look at our the “cleaned up” spreadsheet showing the columns that we need to keep (view example). You should also replace all “#DIV/0!” with a null value and delete any extra sheets.

Third, we are going to identify the census tracts in Los Angeles county that will likely be the best areas to provide foreclosure related services to. We need to set the a criteria for identifying the tracts. What criteria you set is, of course, fairly subjective. When I create criteria like this I tend to select criteria that are restrictive enough to eliminate the majority of tracts but not so restrictive that I the pool of identified tracts is too small to be useful. For the purpose of this tutorial, our criteria will that tracts must have (1) 200 or more units that are owner occupied and (2) have a population that is 70% or more LSC eligible.

Create a new column (IMPACT). Using the Excel “If function” calculate whether each tract meets the criteria discussed above. Our “If Function” will read “=IF(PERLSC<.7,"",IF(H007002<200,"",1))"(view example). This formula will return a one in the IMPACT field if the tract meets our criteria and a null value if it does not.

Create a new column (RIMPACT). Using the Excel “If function” calculate whether each tract meets the criteria discussed above. Our “If Function” will read “=IF(PERLSC<.7,"",IF(H0COLOR<200,"",1))"(view example). This formula will return a one in the RIMPACT field if the tract meets our criteria and a null value if it does not.

Fourth, lets map! Save the Excel file as “Foreclosure Data.” Logon to Dataplace. Select “My Dataplace.” Select “Create New Project.” Title your project and select “Save.” Select “Upload the file from your computer.” Browse to the “Foreclosure Data” and upload it. Select “Extract Tables.” Select “Create Dataset.” Enter the required information. You may leave the default dates or enter todays date. The source of the data is “SF3 2000 Census.”Select “OK.” Select the entry “SFID.” Unselect “Indicator” and select “Part of Region.” Select “Update.” Select “Generate Indicators.” (It will take a few minutes for DataPlace to generate the indicators…be patient.) Lastly, map the indicator of your choice (IMPACT or RIMPACT) and zoom in to the appropriate level so that you can see those Census Tracts were large populations of low-income homeowners live.

• Filed under: Data, Mapping, Housing
• Posted by ElektroMoose | 11:43 am | No Comments »