Civicist

CIVIC TECH NEWS & ANALYSIS

TRACKING

Predicting the west coast civic tech scene; analyzing Twitter for sarcasm; and more.

  • This is civic tech: Microsoft’s Jessica Weare has some intriguing predictions for the west coast civic tech scene in the coming year, starting with the “death of the hackathon” and the rise of long-term partnerships involving volunteer-driven civic innovation.

  • Ash Roughani, the founder of Code for Sacramento, makes a bold and useful proposal for the development of a civic operating system, built by a new kind of public utility.

  • Trump Watch: Continuing the shift in conventional wisdom, here’s Nate Silver of FiveThirtyEight explaining Donald Trump’s staying power: “the [Republican] party isn’t doing much to stop Trump. Instead, it’s making such an effort against [Ted] Cruz.”

  • On Friday, Trump retweeted a white supremacist’s jab at Jeb Bush. This same account has tweeted an animated picture of Trump “dressed as a Nazi, about to gas” Bernie Sanders, Bethany Palma Markus reports for Raw Story.

  • Tech and the presidentials: With a week to go, the top questions Change.org users want asked of the Republican primary presidential candidates on the new ChangePolitics.org platform are on social security, animal testing, and, yes, reforming marijuana laws. The top three aimed at the Democratic candidates are on campaign finance reform (from ex-candidate Lawrence Lessig), animal testing and, yes, reforming marijuana laws.

  • Word games: Longtime readers know that I have regularly railed against claims that such-and-such a tool can provide meaningful “sentiment analysis” of online postings. Language is notoriously squirrelly, after all, and when it comes to political statements, people often employ a variety of tricks—especially sarcasm—to shade or emphasize a point. Well, now come two computer scientist Ph.D.s from Carnegie Mellon with a serious paper suggesting that they have figured out how to detect sarcasm on Twitter with 95 percent accuracy. (Yeah, right.) Their solution is to train their algorithm with a set of tweets marked by their writers with the word “sarcasm” or “sarcastic,” and then to add additional filters that tend to strongly indicate sarcasm. My favorites, in case you want to start confusing your readers, are word unigrams like dare, shocked, clearly, #lol and gasp, and bigrams include you mean, how dare, i’m shocked, i’m sure and at all. No way! (h/t Ian Kar)

  • Speaking of training computers to analyze speech, a team of researchers led by William Li of MIT have figured out the author of unsigned Supreme Court opinions based on comparing the language they use in signed opinions, Adrienne Lafrance reports for The Atlantic.

  • Brave new world: An array of leading civil liberties and internet freedom organizations have written the FCC asking it to start drafting internet privacy rooms “as quickly as possible,” Maria Trujillo reports for The Hill.

  • With the FAA looking closely at regulating nonmilitary drones, lobbying around the technology is heating up on Capitol Hill, Cecilia Kang reports for the New York Times. “Hobby groups are trying to peel back recreational registration rules, while airline pilots are pushing for more mandates that drone makers like DJI and GoPro put safety technology on machines. Amazon and Google, which want to use drones for delivery, are asking permission to test their technology,” she writes.

  • Quartz’s David Panofsky and Jason Karaian built a helicopter-tracking rig to pick up the radio transmissions from aircraft flying into and around Davos and managed to identify 16 private helicopters flying around the area during the elite conference.

  • End game: Erik Wemple of the Washington Post has a savage takedown of Facebook co-founder Chris Hughes’ profligate spending on The New Republic (which he is now trying to sell), featuring multiple office and magazine design renovations, consultants galore, expensive teleconferencing technology that rarely worked, and glamorous A-list parties.