This is a sketch created in the Processing programming language. The bird-like objects re-calculate their heading and steer toward new vectors based on the position of the movable rectangular objects.

Richard Marsico of The New York Law School invited me to be part of the HMDAnalysis project along with fellow NYU classmates and Ballot Boxer partners, Matthew Chmiel and Fernando Cervantes. The focus of the project is an online service to help legal experts, bankers, academics and community activists calculate the complex annual lending data required by the landmark Community Reinvestment Act (CRA) and the equally important Home Mortgage Disclosure Act (HMDA). This project was proposed after an introductory meeting in August 2006 with Professor Marsico, who specializes on enforcement of these laws.

In brief, these Acts require lending institutions to meet the credit needs of the community, including low and moderate income (LMI) neighborhoods, and they require disclosure of information about lending according to race. Since these laws were enacted, banks have made significant increases in loans to LMI communities, in large part because of the efforts of local not-for-profit organizations and other activists. Forcing banks to comply with these laws is complicated and, as a result, people of color are frequently issued loans with terms that are not as favorable as those for whites and borrowers from middle and upper income levels. The government has mandated that all lenders keep, and make public, statistics about their lending amounts, interest rates and how those figures relate to income and race. Although they do make this data available, it is hard to find and even more difficult to manage.

Through his research, Professor Marsico had developed a series of very effective methods for crunching this lending data to calculate discrepancies based on race and to detect predatory lending practices, such as those found in subprime lending. In addition to the fact that the data is vast and hard to locate, Professor Marsico's collection and calculation protocol was complex. Together we refined the process and created an online tool to help users collect, calculate and evaluate this data. We simplified the process by dynamically generating the instructions and calculation tables. This way, users only collect and calculate the data that they require. The HMDAnalysis beta version is currently available to the general public for free and it can support analyses in all 50 states. The project was recently presented by Professor Marsico in Washington.

Various visual and sound designs.

This device, created by Matt Chmiel, Fernando Cervantes and me, is an actual physical boxing dummy wired with a piezo sensor, a microcontroller, an op amp and additional electronics that feed serially into a computer. This configuration allows the dummy to challenge users to a boxing game against a dynamic list of political and media figures. Users interact with a PC-based interface to choose a media figure who they find particularly frustrating, then they enter in their own (the user's) zip code, and their name, a pseudonym or email address. In addition to that information, they have the option of listing the reasons they are interested in boxing this politician or media figure. Then they put on the gloves and start boxing.

Each round begins with a carefully-selected, well-known sound clip of the chosen politician's voice (i.e. George Bush saying "weapons of mass destruction" or Trent Lott remarking positively about Strom Thurman's segregationist presidential platform). In the sound clip, users first hear the crowd hushed and then once the politician speaks, a bell rings and the user can punch away. The outcome of each game determines the output of the device. Winners will automatically transmit emails to the inbox of the opponent they chose, say President Bush, as well as anonymously publish a corresponding blog associated with this website. The data gathered will reveal which political figures inspire the most public attention, in addition to the zip codes churning out the most use. Losers of the game will receive a rematch email. If a desired opponent is not on the list provided on the PC-based screen interface, a nomination form will be offered to add that opponent to the list of nominees. If there is a quorum of nominations, sound clips will be made and fed to the device for future use. There is a web-only version of this project (no boxing dummy) that is currently in proof-of-concept phase online.

In this sample, the sequence of sounds is the crowd being hushed, Trent Lott speaking and the round begins. If you listen carefully, you'll hear the crowd react to particularly forceful punches and the Howard Dean scream to signify the end of that round. Our ideal scenario is to have several of these dummies networked together at major political conventions.

Below are some images taken during the creation of the device in the early winter of 2005. The first two images are of the CD player we initially hacked for this project. We ended up using a PC because we changed the design so that the dummy could use a network to send emails and store data. The next image is of Fernando working out the PIC Basic Pro code, the code we compiled for the microcontroller. The final image is of a version of the circuit. This version was before the op amp was added to boost the piezo sensor signal.