MAPLight Illuminates Connections Between Money, Geography and Politics

By Directions Staff

MAPLight.org, a nonpartisan, nonprofit research group, collaborated with Avencia Inc. to generate an extensive series of maps for MAPLight.org's "Remote Control" report, which shows the connection between money, geography and politics. The maps provide information on the geographic origin of contributions to members of the U.S. House of Representatives. Maps are created to demonstrate the source of funding at the congressional district level of both in-district and out-of-district campaign contributions from 2005-2007. Robert Cheetham, president of Avencia, answered our questions.

Directions Magazine (DM): Can you provide details on the application and mapping tools/software used to create the MAPLight.org maps?

Robert Cheetham (RC):
The maps that illustrate MAPLight.org's Remote Control study were all developed on the desktop using ArcView 9.3. The report required 421 maps "one for each legislator studied" so the primary challenge was development of a process for automating map production. We settled on a combination of a geoprocessing model built using ESRI's ModelBuilder and scripting using Visual Basic for Applications and ArcObjects. The model was used to query the contributions database for information on the given legislator and to update the map display to show those specific contributions. The VBA script took care of updating the relevant text, exporting the document and moving on to the next legislator. The almost 700,000 individual contribution records were geocoded using ArcWeb Services.

DM: Are the data being refreshed on an ongoing basis and who is providing the data? Is this information in the public domain?

RC:
The campaign contribution data were provided to MAPLight.org by the Center for Responsive Politics (http://www.OpenSecrets.org), whose data are based on reports that candidates and contributors are required to file with the Federal Election Commission. Federal and state campaign contribution data are generally available to the public and are updated on a schedule determined by each jurisdiction. While much of these data are now electronic, they frequently contain inconsistencies that require a fair amount of scrubbing to improve the quality. The MAPLight Remote Control report and maps are specific to the timeframe of January 2005 through December 2007. MAPLight.org is working to develop a simpler, interactive version of the maps that will be updated on an ongoing basis.

DM: The maps are static at this time. Are there plans to make them more interactive, or add additional features or data types?

RC:
Currently, the maps only contain information on individual donors to U.S. representatives. Even remaining focused solely on contribution information, there is room to add new data on corporate and PAC donations as well as donations to senators. MAPLight.org also breaks down contribution information based on interest groups and affiliations of individual donors, whose donation geography might be interesting to explore. MAPLight.org is developing a simplified but interactive version of the maps that will be published as a Flash application. This interactive version will also facilitate the regular addition of new contribution data. MAPLight.org expects to make this application available to the public in 2009.

DM: You say that Avencia is also helping congressional campaigns improve fundraising efforts by analyzing donor patterns and demographics, as well as using innovative geographically weighted regression (GWR) tools to predict election outcomes based on a wide range of variables. Do you have any examples of this and can you explain in more detail about GWR? Are you using the new ESRI tools for spatial regression?

RC:
Avencia has developed a suite of GIS and database services specifically for political campaigns, including data integration, statistical data mining techniques, mapping and visualization of political data, and mathematical modeling of donor and voter behavior. For one congressional campaign, Avencia mined large databases of voter and consumer activity and developed a series of "mappable" factors, including voting history, income level, age, race, religion and interest in key issues (economy, environment, education) that might best explain patterns of voter and donor behavior across a district. Using the statistical tools available in ArcGIS, including the ordinary least squares and geographically weighted regression tools, Avencia modeled interactions among these factors and teased out the factors that were most useful in the search for likely donors and voters. The data analysis team at Avencia is also exploring the predictive features of the GWR tool for use in the 2010 election cycle. In addition to congressional campaigns, Avencia has worked with city council and campaigns for state office to create maps that support both campaign logistics and canvassing. Our other election-related work has been with the non-partisan voter protection program run by the non-profit Committee of Seventy (www.seventy.org)

DM: Can you explain more about Cicero and the DecisionTree for planning and prioritizing the get-out-the-vote application? Can you predict where canvassing should be done? Where donors are and how to target advertising?

RC:
Avencia has done donor analysis for both non-profit arts organizations and political campaigns, but these have generally not been interactive applications, but rather off-line analysis and reports. This type of analysis highlights areas that match a particular demographic profile against current donation patterns and looks for geographic areas that fit the profile but are not yet patrons (for arts organizations) or donors.

The Cicero web API can support this kind of analysis as well as other activities such as advocacy and get-out-the-vote (GOTV). Cicero is a web service that features legislative district boundaries, district maps and information about officeholders for the U.S., Canada, Australia and New Zealand. Both the geographic data and the API can be used to stamp lists of addresses with legislative representation information. For both the MAPLight.org project and for a similar state-level project we developed for Common Cause, we have used the Cicero boundary data to assign congressional, state assembly, and state senate district values to each contribution record.

DecisionTree is a web-based solution for geographic prioritization. In GIS technology terms, it is a very fast geoprocessing engine for performing weighted overlay and other raster operations. From a user perspective, it presents a set of factors that might contribute to a particular decision and enables a user to select the ones that are important to him and assign weights to each factor. DecisionTree then uses those weights to produce a "heat map" highlighting the places that best match those factors. It has been used for business siting activities to promote economic development, but can also be applied to election campaigns to geographically prioritize GOTV, canvassing, advertising and other activities by enabling users to weigh factors such as number of voters, high voter turnout, history of voting for particular candidates, etc. There is a sample up and running.


Published Tuesday, January 20th, 2009

Written by Directions Staff



If you liked this article subscribe to our newsletter...stay informed on the latest geospatial technology

© 2016 Directions Media. All Rights Reserved.