DataDriven

Blog

Telling you about the latest in open data and data initiatives in the City of New Orleans.

November 27, 2017

You too can be a “Use Case Truffle Pig”

by Melissa Schigoda, Alex Williamson, and Stephen Primeaux, Office of Performance and Accountability
Filed under: analytics, NOLAlytics

Did you know that truffles (the mushroom kind, not the chocolate kind) are often worth more than their weight in gold? It’s true! Why? Because they are really delicious and really hard to find. Truffles grow nestled among the roots of trees, often deep underground. Lucky for those of us with gourmet tastes, truffle pigs have a keen sense of smell and a natural affinity for rooting in the earth for food that make them experts at locating and extracting these tasty tubers.

Much like truffle pigs hunting for truffles (as founding director Oliver Wise would say), the Office of Performance and Accountability (OPA) team spends a lot of time searching for use cases where we can leverage data science to make changes in city businesses processes. Our ultimate aim is to make local government more efficient and effective. And we need your help!

Last year around this time, we offered all city employees the chance to join us as “use case truffle pigs” and submit their own analytics project ideas. The response was overwhelming!

The following projects were selected and are on track for completion in 2017:

  1. Police retention: Applying machine learning algorithms to historical data on NOPD officers, OPA is developing a model to predict which officers will leave NOPD.  In addition to forecasting future attrition, analysis has identified which factors are most correlated with an officer’s decision to leave NOPD, facilitating a data-driven discussion of possible retention strategies.  This project will continue in 2018.
  2. Traffic cameras and crashes: OPA evaluated whether traffic camera locations experienced fewer crashes than comparison sites. The analysis found that cameras installed from 2008 to 2012 did have a beneficial effect in terms of reducing crashes. That is, road segments with cameras experienced significantly fewer crashes on average than would have otherwise been expected.
  3. EMS ambulance placement:  In partnership with LSU analytics graduate students, OPA developed an algorithm to optimize where ambulances are posted around the city, incorporating daily and weekly patterns in traffic, 911 calls, and ambulance availability.  Under the new protocol, the districts with slowest historical response times – the 4th (West Bank) and 7th (New Orleans East) districts – saw notable improvements. In addition to more equitable response time across districts, there were slight improvements citywide in response times during the night shift.
  4. Blight coin-sorter (phase 2): Updating an earlier analysis, OPA refined a tool that leverages data about the characteristics of a blighted property (e.g. severity of blight, historical significance, levels of fines, strength of market in which the property is located) to recommend to Code Enforcement leadership on whether a property should be sold or demolished. OPA then developed additional criteria to determine which properties are most likely to result in a completed sale, which allows Code Enforcement staff to better allocate scarce resources and prioritize those properties most likely to be brought back into commerce. This scorecard is leading to swifter, more consistent, and more transparent decisions about abatement decisions and is helping code enforcement improve productivity in numbers of properties brought to sale and the proportion of those auctions that lead to successful sales.

Building on the success of these projects, we are again inviting all city employees to submit analytics project ideas. You can sharpen your “use case truffle pig” skills by reading more about our past projects and reviewing the different types of analytics projects.  We look forward to hearing from you and happy hunting!

Photo source: http://vpdm.ca/wp-content/uploads/2015/06/truffle_pig_digital_marketing.jpg