A few years ago, in the height of my workaholism, I took up a hobby. I go to sketchy neighborhoods around L.A. and hang out with dogs I don’t know.
I have a long history of adopting and fostering shelter dogs, often getting them out on their “euth dates.” With almost 4 million homeless animals losing their lives every year in the United States, I wanted to do something more than robotically write donation checks. So I go to the shelter, choose the most hard-to-place dogs, videotape them, and post the videos on social media. It turns out that 93 percent of the dogs I video find homes.
How do I know this? From my data! I keep a running log of the dogs I video, their breeds, their ages, and the reason they enter the shelter in the first place. Though it might not be hugely voluminous—I’ve videotaped hundreds of dogs, not tens of millions—if you’re a dog hours away from being put to sleep it’s “big data” to you.
Some of my findings are no surprise. Anyone who’s been in an animal shelter knows that the vast majority of dogs who need homes are pit bulls (according to my data it’s 46 percent), followed by Chihuahuas (23 percent). Then come German Shepherds, Chows, Rottweilers, and variations of all of the above. While mixed breed dogs figure prominently, 38 percent of the dogs I meet in the shelter are purebreds.
But the potential for this data transcends my Facebook videos. What if the politicians who govern public shelters were data-driven?
We could enrich data about dogs and adoption rates, discovering demographic clusters of owners who surrender their dogs, and where they live. This could justify funding for mobile spay and neuter clinics in those neighborhoods. We could pinpoint patterns of abuse, and use that to focus law enforcement efforts, or reward informants for exposing dog-fighting rings. We could fund intervention programs to help low-income families find housing and low-cost veterinary services, allowing them to keep their pets.;