Shailesh Prakash, the chief information officer for The Washington Post, was in techie heaven as he told a crowded Northwestern University lecture room of geeks and newsies about Hadoop, Spark, MongoDB, Druid, Kafka, HBase, AWS and other Jeff Bezos-inspired gambits married to Post journalism.
They all appear to be leaving much of a geriatric, scared and under-resourced newspaper industry in the technological dust.
It also helps explain why the paper’s domestic digital traffic, which was 30 million in October 2013, is approaching 100 million unique visitors a month, or roughly one in three Americans. And how the Post segments advertising to individual consumers, quickly identifies underperforming stories and tries new headlines (“content variation testing”); marries ads and related content; uses predictive algorithms rather than a slavish adherence to click rates to assess stories’ prospects; uses a “Virality Oracle” bot to help predict popularity and vitality of content; and tests whether machines can do better at writing headlines.
For example, he showed a comparison of a human-written headline and two done by machine:
Human: Theresa May vows to “lead Britain forward” despite staggering election blow
Machine: British prime minister to stay in power
Machine: Diminished May vows to stay course
The obvious takeaway at the “Computation + Journalism Symposium” — put on by Northwestern’s school of engineering, office of research and Medill Journalism, as well as Mozilla and Google News Lab — is that machines have a long way to go but that they’re making progress. But there was a lot more, all underscoring why the Post
Read more here: https://www.poynter.org/news/washington-posts-innovation-sage-future-coming-along-nicely