Resistance Radio: Data for Democracy – Racial Disparities in Police Traffic Stops

Data for Democracy is a volunteer collective of data scientists throughout the country that emerged since the election of Donald Trump. It has local chapters taking a look at civic data in their local communities. Sam Zhang and Jeffrey Moore are members of the Boulder chapter of Data for Democracy and they’ve been digging into some data on policing that has been released by Stanford University.

Zhang says that dataset shows that traffic stops and searches by police have dropped by nearly half after the legalization of marijuana, however there are still significant racial disparities.

“Latino folk are stopped twice as often as white folk and contraband is found on them 30% to 50% less often, so you can’t justify stopping a population more frequently if you’re finding contraband on them less frequently. Until this data was released, this was only a hunch – a very obvious hunch to people who are affected by it – but a hunch nevertheless.”

Denver analysis:

The data analysts grouped together all of the state police stops made in Denver County, Douglas County, Adams County, Arapahoe County, Jefferson County, Boulder County, and Clear Creek County.

The curves in the graph below represents the group’s beliefs about the rates at which police find contraband when they search drivers of each race. Zhang says they are confident that Hispanic drivers are searched less successfully than White drivers. “We also find that searches of Hispanic drivers tend to be less successful than searches of White drivers somewhere between 33-48%.”

Boulder analysis:

Looking at Boulder specifically, Zhang says they have less data so their inferences are less certain (they overlap more), “but we are still certain that Hispanic folk have a lower search success rate than White folk.”

The analysts didn’t break apart the data from before and after marijuana was legalized, though the Stanford research team did, and found that biases persisted afterwards in Colorado: http://www.nbcnews.com/news/us-news/police-searches-drop-dramatically-states-legalized-marijuana-n776146

The code in general is available online at https://github.com/samzhang111/co-police-stops

The Boulder Chapter of Data for Democracy has also been working on “social good” projects with a data science angle, like expanding the collection of regulatory disclosures at the AirWaterGas project at CU, and looking at city data with the city’s Chief Innovation & Analytics Officer.

Zhang says for future projects, he would like to analysis data of complaints made to the COGCC (Colorado Oil and Gas Conservation Commission.) “When you get fracked, when your neighborhood gets a well in your backyard pretty much, and you have noise pollution and you have weird issues with your drinking water and your house keeps shaking, what the officials will tell you to do is “we don’t have much to do for you, but you can submit a complaint to the COGCC”, but so far I haven’t seen very many journalistic exposes of this, but it seems like a lot of times the complaints just get referred to the operator “oh your waters bad…well tell Crestone about it and they’ll look at it for you” and it’s considered resolved…”

Moore says he would like to do some data analysis on government transparency. “I am personally interested in just transparency at the local city council level, what’s being done in the city, what’s being discussed at our meetings and how our representatives are best representing us and making sure people know exactly what is being discussed and if the issues they’re most interested and concerned about are being raised.”

Podcast: Play in new window | Download