Killed it with a #Killfie: Journey from an Idea to a Global Media Phenomenon

31,000+ likes, 34,000+ shares, 1,000+ Tweets!

Most research goes through some natural phases; formulating the problem statement, collecting and analyzing data, submitting a research paper to a conference, writing a technical report, and then hoping the paper will get accepted at the conference and the work will be appreciated/acknowledged by the community (happily ever after!). I had never imagined that one such research topic, which went through some initial natural phases, will take such an interesting turn at some point and receive such an overwhelming amount of attention!

A lot has been said and written about our recent work (you can infer that from the title, and see ‘Who is talking about this research’) both in the technical community and press. I want to share my behind-the-scenes experience of going through this amazing phase of research – when it gets hard to count the number of mentions about your work returned by a quick google search! A news article about someone dying just after taking a selfie was posted on the Precog mailing list on June 2, 2016. Definitely not a conventional cause of death, this disturbing news made some members of the group to dig into the what, how and why of selfie deaths around the world. It was just a small idea that we started working on, discussions trickled, and some compelling observations followed. All culminated into a well written paper, submitted to a conference and the technical report going online on Arxiv on 7th November.

The report was first picked up by Sun UK news and some twitter handles like VickiTurk  on 9th November and what followed was a whirlwind of news articles and technology blogs across the globe, and across all media. It had become a sensation! It seemed to have touched all time zones from California (GMT-8) to New Zealand (GMT+13). The news buzz peaked on 18 November when three of us, Hemank Lamba, Megha Arora and myself, went on a spree of giving interviews. We had news reporters wanting answers over email, phone and skype, following up with us through the day. Here is what 18th November entailed for all three of us:

  • 0700 hrs IST: Call with CBC Canada, me sitting in IIT Kharagpur guest house and Hemank and Megha taking the call from Pittsburgh
  • 1000 hrs IST: Call with BBC UK Radio, I was taking it alone from in IIT Kharagpur, CSE department
  • 18 00 hrs EST: BBC World TV News, Hemank took this alone from Pittsburgh
  • 2130 hrs IST: NBC US, I took the call during my transit from Delhi Airport to IIITD
  • 2330 hrs IST: CMU, Hemank and Megha sitting in Gates building in CMU and I was at home in Delhi

This does not include the 25+ unique emails that we probably sent out answering questions or fixing timeslots for more interviews. While 3 of us were engrossed in this craziness, Mayank Vachher, Varun Bharadhwaj, and Divyansh Agarwal had their hands full providing backend support, collating the hits that we were getting in news and social media, and getting more specific insights from the data which the reporters were interested in. Three of them played an integral role in ensuring that we had a smooth run. In the meantime, we also had our group meetings to discuss the feedback that we are getting from people around the world.

The news about this research has been spreading across many newspapers, and online social networks like Facebook, Twitter, and Youtube. As of this moment, the following numbers summarize the traction garnered by this research:

  • Total Articles written (unique ones): 160
  • Total Facebook posts: 100+
  • Total Facebook likes: 32,108
  • Total Facebook shares (shares of the articles + posts): 33,937
  • Total Facebook comments: 2,795
  • Total Twitter tweets: 1000+
  • Total Twitter RTs (of all the above tweets): 1075
  • Total videos created on the project: 15
  • Radio interviews: 11
  • TV interviews: 2
  • Total number of requests for the dataset: 6

Below is a tag cloud capturing all the major news agencies which featured our work and the work was featured in 17 different languages.

Lessons learned through this media frenzy:

  • For a research to get popular, the topic has to be relevant to ‘people’
  • Reporters ask interesting ‘research’ questions, be prepared
  • Sociological/psychological studies around ‘who’ and ‘why’ of the research are important
  • Feedback from people is helpful in identifying potential issues in the research
  • Having captivating titles for the paper helps

Below is an infographic capturing the research work.

For those interested in knowing more about this research, here are some useful links:

Published by

PK

Associate Professor @ IIIT-Delhi http://precog.iiitd.edu.in