Just for curiosity's sake, how is this dataset populated? Is ThriftDB just scraping HN? Is there some kind of back-end DB feed?