Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Jochen Spangenberg - Sourcing stories on social media


Published on

Jochen Spangenberg, innovation manager at Deutsche Welle and the Reveal project, gave his advice on sourcing stories on social media at the recent news:rewired digital journalism conference

Published in: News & Politics
  • Be the first to comment

  • Be the first to like this

Jochen Spangenberg - Sourcing stories on social media

  1. 1. Verification of UGC / Eyewitness Media:Verification of UGC / Eyewitness Media: Challenges & Approaches Jochen Spangenberg Innovation Manager // Deutsche Welle & REVEAL Project & Twitter: @RevealEU & @jospang & @dw_innovation London, 16 July 2015
  2. 2. Note: this is a slightly adapted version of the presentation held at news:rewiredNote: this is a slightly adapted version of the presentation held at news:rewired on 16 July 2015 in London It is made available to participants and others interested in the topic verification of eyewitness media (or UGC, as it is often called, too) Copyright notice: Copyrights, trademarks, logos etc remain with respective copyright holders. Usage of logos, images, screenshots etc in this presentation is for non-commercial demonstration purposes only. Respective rules for usage apply.
  3. 3. ??? Source: Georg Mittenecker (CC BY-SA 2.5)
  4. 4. Sources of all images: see for respective copyright holders Finding out how journalists deal with eyewitness media: issues, challenges, wishes, concerns (extracts of selected interviews on
  5. 5. Wishful thinking … Source: Scientology „e-meter“. Image by Salimfadhley. (CC BY-SA 3.0)
  6. 6. Reality check Verification of UGC / eyewitness media • can be laborious• can be laborious • can be time-consuming • requires particular (new) skill sets • Does not (yet) follow industry-wide rules / practices
  7. 7. ButBut
  8. 8. Source: Janis Krums, who took the photo in Jan 2009 on his iPhone and posted on TwitPic / shared via Twitter, from where it was picked up by numerous news outlets. See!/jkrums/status/1121915133) Source: images circulating on Twitter, claiming to show crashed Germanwings flight 9525 Source (of first manipulation): unknown
  9. 9. So what to do?So what to do?
  10. 10. Dealing with (Verification of) UGC / Eyewitness Media Media organisations: be a facilitator Journalists: - value of established (and useful) journalistic practices / skills ... - knowledge of useful tools (usage, benefits / shortcomings) ... Sources: Screenshots of Suncalc, Google Maps & Translate, Tweetdeck, Followerwonk, Mentionmapp, Topsy, Storyful Pro, Jeffrey’s Exif Viewer, Panoramio, Pipl
  11. 11. How?How?
  12. 12. • Is it really what it’s supposed to be? • Is it “too good to be true”? Source: that-really-a-picture-of-hurricane-sandy Source:
  13. 13. Source: screenshot of aSource: screenshot of a Google Reverse Image Search Source: screenshot of a Tineye image search Source: screenshot of a Wolfram Alpha weather check for New Jersey on 29 Oct 2012 (supposed date of photo with shark in New Jersey) Source: screenshot of an Exif data check performed with Jeffrey‘s Exif Viewer.
  14. 14. The Big W’sThe Big W’s
  15. 15. What? Source : blog.
  16. 16. Who? Source: screenshots of Sources: screenshots of Twitter Profiles Source: screenshots of mentionmapp analysis followerwonk analysis
  17. 17. Where? Source: Panoramio screenshot Source: Flickr, by sarflondondunc, (CC BY-NC-ND 2.0) Source: Panoramio screenshot Source: Flickr, by Paul Stein, (CC BY-SA 2.0) Source: Flickr, by s1lang, (CC BY-NC-ND 2.0)
  18. 18. When? Source: YouTube metadata screenshot Source: Flickr, by Petr Dosek, (CC BY-SA 2.0)
  19. 19. Why? Source : Screenshot of Twitter profiles of IDF and Al Qassam Brigades
  20. 20. Source: REVEAL / Linda Rath-Wiggins
  21. 21. The REVEAL approach • Can we teach (develop) algorithms that can distinguish truths from lies / facts from manipulations (assuming that there is such a thing called ”truth“)? ContributorC C C ontributor ontent ontext
  22. 22. The REVEAL approach R&D dealing with, among others Contributor • Who contributed what? • What did they do in the past? • Who follows him/her? With whom do they interact? How? What’s their reputation? Tustworthiness? • How long did social accounts exist? What can be derived from it? • Links to profile pages, affiliates, contact data? • ...
  23. 23. The REVEAL approach R&D dealing with, among others Content • Image analysis (checks, similarity search, manipulations, comparisons) • Text analysis (stylometry, accuracy, comparisons) • (Rudimentary video analysis > not in focus) • ...
  24. 24. The REVEAL approach R&D dealing with, among others Context • Who says what how about the same / related topic, and what is being said how? • Who are the people contributing? What did they do previously? How did they act/react? Can patterns be detected? • How can information be connected sensibly and beneficially with other resources. (E.g.: correlate Social Media activities with information from disaster management / relief agencies / natural catastrophe centres etc) • ...
  25. 25. User-driven GUI/interface development Source: REVEAL
  26. 26. Functionalities Source: REVEAL /
  27. 27. Work on 41 functionalities (we call them modalities) Exemplary modality Geospatial, social and topical context information (developed by IT Innovation) Source: REVEAL / IT Innovation
  28. 28. Some final remarks
  29. 29. Direct contact with source (but there are exceptions) Source: plenty.r. (Flickr, CC BY-SA 2.0)
  30. 30. Source protection / security of sources Source: Dina Regine (Flickr, CC BY-SA 2.0)
  31. 31. Crediting, sourcing and getting permission Source: Extract from a ZDF TV Programme / news report
  32. 32. Ethical aspects Source: Jordi Mir, on YouTube here
  33. 33. Our business Source: Flickr, Terry Johnston , CC BY 2.0 ”Getting it wrong once can be more powerful than getting it right 1000 times“
  34. 34. Useful resources
  35. 35. Useful resources Source:
  36. 36. Useful resources Source: screenshots of respective publications (all available online)
  37. 37. Useful resources (selection) • Bellingcat – collaborative investigations, initiated by Eliot Higgins • Verification Junkie – directory of verification tools by Josh Sterns • Eyewitness Media Hub – legal, ethical and logistic issues (Claire Wardle, Sam Dubberley, Jenni Sargent, Pete Brown) • Research Clinic by Paul Myers – collection of research links and articles • Citizen Evidence Lab – by Amnesty International, guidelines for verifying footage in videos • Emergent – collaborative debunking (no longer in full operation) , by Craig Silverman •’s Malachy Browne’s “pocket guide on verifying details of a video” and other useful guidelines • Storyful (paid verification service) and Storyful Open Newsroom (collaborative verification platform) • Craig Silverman’s Regret the Error on Poynter • Work of Meedan (i.e. Checkdesk) / Tom Trewinnnard et al. • Witness Blog & website – fighting for human rights / against human rights abuse • Authenticating Open Source Video – a Witness tipsheet • Link tips for Social Media research, by Konrad Weber • Correctiv case study on downing of flight MH17 • WAN-INFRA article by Julie Posetti and Craig Silverman on newsroom / verification issues • Resources of the BBC Academy, such as this contribution by Trushar Barot and this one by Alex Murray • ....
  38. 38. Useful resources / great case studies & advice
  39. 39. Thank you! Contact Jochen Spangenberg DW Innovation Berlin Tel: +49 (0)30 4646 5604, +49 (0)172 261 2315 Twitter: @jospang, @revealeu, @dw_innovation &