Payday California

Payday California

Organisation: The Center for Investigative Reporting (United States)

Publication Date: 04/07/2015



Open data award. After California taxpayers discovered the tiny town of Bell had been paying enormous and illegal salaries to officials there, many people asked: How did we miss this for so long? That’s when The Center for Investigative Reporting set out to create the most comprehensive database in the country of local government salaries. Although these salaries are public records, most taxpayers know little about whether the paychecks for city and county officials are fair. No statewide standards govern how local pay is set, leaving the public in the dark about whether their city managers, for example, are paid appropriately for the job and the community. With Payday California, CIR skillfully put into context the $40 billion a year that California cities and counties spend on their employees. Below is the link to the main app. The stories and methodologies are here:

Technologies used for this project:

For the analysis, we sought to identify the factors that influence public officials’ compensation. Using the state’s data from all 482 cities and 58 counties, we did a linear regression analysis and found that much of the variation in top officials’ compensation could be explained by communities’ population size and median rent. We then looked into outliers and examined why they were paid so differently than people in comparable jobs in similar places. We used Excel, SPSS and PostgreSQL. For the app we used several technologies. Payday California is a significant upgrade in features and user experience from previous salary databases. Most salary databases produced by journalists offer very simple functionality, limited mostly to searching and sorting. This leads users to focus primarily on individuals’ pay and highly paid employees. While there are certainly stories to be told through this focus, Payday California was built specifically to create context about a government entity’s pay. Each city or county has its own page, which shows charts that detail the distribution of pay, from lowest to highest, and lets users compare the city or county to statewide averages, or to previous years in the same entity. Each city also includes a breakdown of how departments spend their payroll budget, which gives users a better sense of what a government’s priorities are, both in staff levels and money spent on salaries. We also needed several language processing tools to make the site work, given the scale and messiness of the data. Communities reported every imaginable (and many unfathomable) variations in the names of departments and position titles, making it very difficult to categorize and compare them. We used a naive Bayes classifier to group our departments into general categories, such as “public safety” and “central government administration,” which were used to color code our department charts into something readable. To join city and county names with U.S. Census American Community Survey demographic data and Census map boundaries, we used the Python library fuzzywuzzy to find the best match for city names. On individual salary pages, Payday California includes charts to show how a given employee’s pay compares with that of others in his or her community or department. It also allows users to compare the employee’s pay with the pay of other employees who have similar job titles across the state. That similarity searching uses Solr’s MoreLikeThis fuzzy searching to return results. Users can customize what they want to see as a “similar position,” which adjusts the charts dynamically. We built a highly customizable Javascript widget (still in beta) for other news organizations to embed California salary data on their own sites, and also a tool to help users build their own widget using a simple questionnaire.
Follow this project

Comments (0)

You have to be connected to contribute

You have to be connected to follow

Leave this project and no longer be informed about this project

By joining this project, you will be informed by email when an update or a new contribution is posted on the website.

Thank you for your active participation !

The GEN Community Team