FREE ELECTRONIC LIBRARY - Theses, dissertations, documentation

Pages:     | 1 |   ...   | 31 | 32 || 34 | 35 |

«Urban Problems and sPatial methods VolUme 17, nUmber 1 • 2015 U.S. Department of Housing and Urban Development | Office of Policy Development and ...»

-- [ Page 33 ] --

For Central Falls, Pawtucket, and Woonsocket, addresses from the emergency 911 point shapefile were standardized and matched to parcel records from each city’s taxroll records. This process resulted in separate files specific to each municipality. The geocoded shapefile associated with the emergency 911 addresses served as a good starting point for identifying all the properties in each municipality. Inaccuracies and differences in completeness in these data continue across municipalities, however, and so we would not suggest relying solely on this source. For example, some cities’ records include parcel identifiers in the file, but many do not. Furthermore, because the shapefiles are point data, rather than polygons, they provide no sense of the size and shape of property lines in relation to one another in the city.

The next step in our data preparation entailed actively looking for multiunit properties that were likely to have multiple addresses. Within each city, we selected the tax class that corresponds with

–  –  –

two- to five-family properties and, for each address, added 2 and subtracted 2 from the street number to create fields representing what would be the next address to the left or right of the property.

If those addresses matched an existing address in the data, we discarded them, but, if they did not, we kept them as potential matches to expand the coverage of the MLT. If address numbers were recorded as ranges, we split them up to create multiple address records from the original.

Adding the municipality-specific datasets into one master file was the final step. To avoid any loss of data that could arise if plat and lot numbers overlapped in different cities, we created a city-specific parcel identifier field called “CKEY.” The identifier field concatenated a two-letter abbreviation of the city name—CF (Central Falls), PA (Pawtucket), PR (Providence), and WN (Woonsocket)—and the parcel identifier code: the plat and lot numbers. For example, if the plat and lot numbers for a property in Providence were 1 and 1, the CKEY would have been PR 1-1.

The most basic version of the MLT file contains all residential addresses, the municipality, and the corresponding CKEY, with all text in uppercase lettering to ensure standardization. The basic structure of the file is provided in the appendix. Variables that can be included as needed in the MLT include address number, tax class category, property type, and year built—essentially, any descriptive attribute that is unlikely to change regularly. Because the MLT is focused on linking addresses to parcels, each address within a city should be listed only once and addresses should not match to more than one plat and lot. Parcel identifiers can have duplicates, however, because when they match to more than one address, they are listed in separate rows in the table. The MLT created for the Lead Technical Study currently contains more than 100,000 address-to-parcel linkage records, with about 65 percent of those in Providence.

Providence MLT Online Tool A secondary outcome of the MLT effort was the creation of an online tool that enables the user to import addresses to tag with parcel identifiers or download lists of addresses and parcels. It also enables registered users to modify records when new information is available or errors are found.

See exhibit 1 for a visual example of how the MLT acts as a crosswalk between addresses and parcels. This online version was created using Leaflet, an open-source resource for interactive maps.

Although the tool is currently available for Providence only and is a private website that requires a login, the concept is replicable.

–  –  –

Exhibit 1 Providence Master Lookup Table Note: This screenshot of the master lookup table website provides an example of the interactive nature of the online tool.

By clicking on records BW and BX, a pin shows up on the map. By clicking on the parcel, a box in the top right corner of the screen provides details about the property type, property owner (suppressed), and addresses associated with the particular parcel identifier.

Connecting the Data After the MLT was complete, the next steps involved gathering the various datasets necessary to answer the research questions and then linking them together.

Study Datasets Using the MLT for the four cities in the analysis, we linked city tax assessors’ datasets, two lead compliance certificate datasets (from the Rhode Island Department of Health [HEALTH] and Rhode Island Housing Resources Commission), blood lead screening surveillance data from HEALTH, and foreclosure deeds datasets. Taxroll data provided the details needed to identify if a property was subject to the law, including year of construction and the owner’s address. The lead compliance certificate records were integral to the policy evaluation, which aimed to assess whether outcomes differed based on having a certificate. Lead screening records provided the primary outcome variable—blood lead level screening results—and the address at the time of each screening. Rhode Island law mandates that healthcare providers screen all children for lead twice by age 3 and report the results, so the lead screening dataset covers most young children in the state. Foreclosure deeds allowed for compliance and lead exposure comparisons at properties being considered for possible foreclosures.

Cityscape 211Sylvaria, Cigna, and Lee

Each dataset used different formatting of address, unit, and property records. With the exception of the taxroll and foreclosure records, parcel identifier codes often either contained obvious errors or were completely missing in the datasets analyzed. The lead compliance certificate datasets had errors and missing records for the parcel identifier fields. The screening surveillance data records included only patient address fields, which were subject to data entry errors. In some cases, addresses included post office box addresses, which are not useful for this work because they do not indicate the physical address of residence and thus were excluded. Standardizing the addresses was essential.

Preparation for Matching To prepare the existing address-level datasets, we ensured that any addresses conformed to the same format as the MLT. Geocoding the addresses (we used ArcGIS software) is a good way to begin the process, because it can correct some spelling errors or other inconsistencies automatically.

Further editing required reviewing all addresses for errors, stripping any unit designations out of the addresses (apartment numbers, floor numbers, and so on), making sure public housing and similar complexes were in the same format, and standardizing common abbreviations such as “N” instead of “North.” To match the format of the MLT, we also ensured that any text was all upper case. We used statistical software to clean up the datasets after geocoding, which systematized the process, because the syntax can be adapted to apply to each dataset and reused when new data need to be processed. We used IBM SPSS Statistics software, but other statistical packages would serve the same purpose. Each type of data we prepared for analysis had one file that contained address and municipality fields.

Matching When the separate data files were fully prepared, we could match the addresses in the datasets for analysis to the addresses in the MLT. To avoid any complications with duplicate addresses across municipalities, the files were matched based on two variables: address and municipality. Matching one file per city at a time to the MLT records for that city could also avoid duplicative addresses.

By processing all the address-level data through the MLT, we ensured that the addresses in each of the various datasets correspond to the same properties. Matching to the MLT resulted in each file containing the CKEY parcel identifier field.

After all the files contained a CKEY field, one main property-level analysis file could be created.

The taxroll file at the CKEY level served as the base file to match the other property data. For the purpose of the Lead Technical Study, a subset of residential properties built before 1978 was selected—one- to five-family properties, apartments, and mixed-use properties—based on the corresponding codes in the taxroll data. Properties were then classified as either owner- or nonowneroccupied. We aggregated the address-level taxroll, foreclosure, and lead certificate records to the parcel level using the CKEY variable and matched the files to the taxroll. New fields indicated if a property had lead certificate records, any foreclosures, or was exempt from the law. See exhibit 2 for example match rates based on the foreclosure analysis from 2005 through 2009.

–  –  –

Note: The match rates report the percentage of properties in the dataset from which they originate that matched the properties in our pre-1978 taxroll records.

The way we handled the lead screening data differed slightly depending on the unit of analysis. To prepare the lead screening data after matching it with the MLT, we analyzed one record for each child. If a child matched to more than one property, we kept one record per child per property. For the child-level analysis, we matched the property-level file of taxroll, foreclosure, and lead certificates data to the lead screening file to describe the property where the child was living and focused on the first result per child per property. For the property-level analysis, we aggregated the number of children who had screening records at a given address and identified whether one or more children’s maximum test results were considered elevated. Thus, conducting analysis with these methods allowed for flexibility at the aggregation level.

Analyzing the Data The crosswalk provided by the MLT helped create datasets that could be analyzed with relative ease at various levels of aggregation. For the analysis that led to the American Journal of Public Health article (Rogers et al., 2014), we focused mainly on outcomes at the child level and could account for children who lived in different properties or multiple children who lived at one property. We identified whether children who lived in properties that became compliant had declines in blood

Cityscape 213Sylvaria, Cigna, and Lee

lead levels, and we described the burden of lead exposure in exempt properties. For work outside the scope of the journal article, we investigated outcomes at the property level to describe how many properties housed one or more lead-exposed children, comparing lead exposure in compliant and noncompliant properties as well as between exempt and nonexempt properties. These property-based findings were presented to stakeholders and at conferences. Finally, we analyzed children’s lead exposure by whether the property where they lived had been foreclosed on within a certain amount of time. When new data become available, these methods can be repeated to keep stakeholders updated on the status of associations of interest.

Discussion Investing the effort to match data with the methods described in this article has numerous advantages, especially in urban communities with high proportions of multifamily properties. In many cases, administrative datasets that are relevant to housing and health will have address records, but not parcel identifiers. A data crosswalk such as the MLT provides not only a way to link disparate datasets but also rich layers of information to analyze. Without preparing a comprehensive propertylevel lookup table, a much higher proportion of data would be lost because of nonmatching. In addition, without translating data from address- to property-level status, count, proportion, or density calculations for a given area could mislead readers. The data are particularly misleading at smaller geography levels, such as census blocks and neighborhoods; the differences between property and address information can be meaningful.

The techniques employed to create the MLT and use it as a data-matching system could benefit researchers conducting analysis at the property level in other cities in the absence of established integrated data systems. The tool supported the evaluation of outcomes associated with housing policies—work that would have otherwise been unfeasible given the data landscape. Although the effort associated with having to conduct this work one municipality at a time could make a statewide analysis burdensome, the ability to target at-risk communities and, in some cases, at-risk properties has been valuable to stakeholders in Rhode Island. In summary, ensuring that linkages between property data are accurate and meaningful can lead to meaningful results. Those robust analyses can, in turn, guide policymakers to evaluate housing-related policies more effectively.

–  –  –

Acknowledgments The authors thank Jim Lucht, Kimberly Pierson, and all who were involved with the creation of the original master lookup table. They also thank Michelle Rogers for her input on this piece and her great work on the analysis, as well as Patrick Vivier, Ryan Kelly, Robert Vanderslice, and all those people who assisted with the lead technical study work thus far.

An award from the U.S. Department of Housing and Urban Development provided funding that supported the work that provided the basis for this publication. The substance and findings of the work are dedicated to the public. The authors are solely responsible for the accuracy of the statements and interpretations contained in this publication. Such interpretations do not necessarily reflect the views of the U.S. Department of Housing and Urban Development or the federal government.

Authors Alyssa J. Sylvaria is a health policy and information specialist at The Providence Plan.

Jessica Cigna is the research and policy director at HousingWorks RI at Roger Williams University.

Rebecca Lee is the director of the Information Group at The Providence Plan.

–  –  –

References Healthy Housing Collaborative. 2012. “Healthy Housing Data Book.” http://www.health.ri.gov/ publications/databooks/2012HealthyHousing.pdf.

Rhode Island Geographic Information System. 2014. “Facilities and Structures.” http://www.edc.


Pages:     | 1 |   ...   | 31 | 32 || 34 | 35 |

Similar works:

«95 THEORETICAL RESEARCH Contemporary Social Work: an Analysis based on the federal budget Janaina Albuquerque de Camargo Schmidt Mossicléía Mendes da Silva Universidade Estadual do Rio de Janeiro (UERJ) Universidade Estadual do Rio de Janeiro (UERJ) Contemporary Social Work: an Analysis based on the federal budget Abstract: This study presents a critical reflection about Social Assistance policy, based on considerations about funds allocated to this policy in the federal budget highlighting...»

«ISSN (Print) : 2320 – 9798 ISSN (Online): 2320 – 9801 International Journal of Innovative Research in Computer and Communication Engineering Vol. 1, Issue 2, April 2013 Distributed Firewall Application for Policy Management and Network Security Manila Bohra1, Laghvi Aloria2, Neha Gupta3 B. Tech (6th sem) Student, Rajasthan Technical University, Kota, Rajastan, India1 B. Tech (6th sem) Student, Rajasthan Technical University, Kota, Rajastan, India2 M.Tech Student, Rajasthan Technical...»

«NOT IN THE FINE PRINT: RECOMMENDED CHANGES TO LIFE INSURANCE POLICY DISCLOSURES REGARDING RETAINED ASSET ACCOUNTS MICHAEL A. BARRESE* I. INTRODUCTION Tom is the primary breadwinner of his family. In order to protect his wife and children financially in the event that he passes away, he goes online and researches life insurance policies. After becoming familiar with the different forms of life insurance, Tom purchases a $250,000 life insurance policy from a large insurance company. When he...»

«centre for analysis of risk and regulation An ESRC Research Centre Analyzing Public Management Policy Cycles in the European Commission: Oversight of Budget Control and the Integrated Internal Control Framework Michael Barzelay, Roger Levy and Antonio Martin Porras Gomez DISCUSSION PAPER NO: 65 DATE: August 2010 Analyzing Public Management Policy Cycles in the European Commission: Oversight of Budget Control and the Integrated Internal Control Framework Michael Barzelay, Roger Levy and Antonio...»


«A System Perspective for Policy Analysis and Understanding: The Policy Process Networks A System Perspective for Policy Analysis and Understanding: The Policy Process Networks Yael Parag Visiting Research Fellow in the Environmental Change Institute (ECI), Oxford University Centre for The Environment (OUCE) Abstract This paper introduces the Policy Process Networks (PPN) as a new framework for policy understanding and analysis. PPN combines the strengths of the Policy Cycle and the Policy...»

«Youth Policy: What Works and What Doesn’t? A Review of Youth Policy Models From Canada and Other Jurisdictions February 2008 Youth Policy: What Works and What Doesn’t A report of United Way Toronto February 2008 Prepared by: Kamara Jeffrey United Way Project Staff: Diane Dyson Kathy Gallagher-Ross Michelle Smith Ming-Young Tam Peter Alexander United Way Toronto thanks and acknowledges the following individuals for their contribution of time and expertise: Angela Carr, Government of New...»

«Australia's Involvement in the Vietnam War, the Political Dimension Part 1 © Brian Ross, 1995 Introduction This is the second post promised analysing why Australia entered the Vietnam War. American readers should be warned that because it looks primarily at the domestic political scene in Australia at the time, it does as a consequence refer to characters and events which most of you will not be aware of. However, I have included a short preface, attempting to identify most of the major...»

«NPSIA Works: Career Futures CANADIAN PUBLIC SERVICE – FEDERAL AND PROVINCIAL Description of the Field In addition to international jobs found within DFAIT, almost every federal department has an 'International Division' which relates to international and other national organizations in its subject areas. Activities in this division may include researching another country’s experiences, exchanging information with a department’s counterpart in another country, preparing policy papers on...»

«Electronic and Information Resources Accessibility Policy and Procedures Manual Table of Contents Chapter 1: Overview 1.1 Purpose and Scope 1.2 Application 1.3 Compliance 1.4 Lack of Commercial Availability Chapter 2: Website Accessibility and Usability 2.1 Texas Web Accessibility Standards 2.2 Linking and Indexing State Websites Chapter 3: Functional Performance Criteria for Websites and Applications 3.1 Overview 3.2 Criteria to Assist Users with Visual Impairment 3.3 Criteria to Assist Users...»

«CANADA PART TIME EMPLOYEE POLICY MANUAL July, 2008 TABLE OF CONTENTS TOPIC PAGE INTRODUCTION • Part-time Associate Policy Manual EMPLOYMENT POLICIES • CROSSMARK CANADA and Diversity • Diversity Direction • Key Objectives • Equal Employment Opportunity Policy • Employment Qualifications • Immigration Requirements • Associate Status • Introduction Period • Regular Full-Time Status • Part-time Associates • Temporary Associates • Termination for Reasons other than Cause...»

«United States Department of the Interior June 2008 DEPARTMENT OF THE INTERIOR INTEGRATED CHARGE CARD POLICY MANUAL This page intentionally leftblank. Table of Contents, 1.2 CREDITW ORTHINESS,, Credit Scores 1.4 ROL AND RESPONSiBILITIES 1.4.5 Office of Inspector General (O–  –  – The Department of the Interior's integrated charge card for travel, purchase, and fleet is Bank of America's MasterCard. This card is specifically designed with the United States ofAmerica printed next...»

<<  HOME   |    CONTACTS
2016 www.theses.xlibx.info - Theses, dissertations, documentation

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.