It says in the article they employed a commercial data set that is not composed of widely available public data in creating this app. Also, it lists a lot more information about people than whether or not they are christians.

I also returned totally accurate results using the exact same query. I would really like to know what is going on here. This is a common complaint with some people using DDG, that the results are poor, but I consistently have as good if not better results than using Google.

Hell, if you’re scanning the trash for cans to separate out and recycle, why not just scan all the rest of the trash and figure out whatever information you can from there. You could realistically scan all the trash and log every identifiable piece while only removing cans and logging all the data. Don’t know how much valuable information one could pull from this data that isn’t already available through sales data, but it’s an interesting concept to mull over at least.