Dataconomy
  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Glossary
    • Whitepapers
  • Newsletter
  • + More
    • Conversations
    • Events
    • About
      • About
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
Subscribe
No Result
View All Result
  • AI
  • Tech
  • Cybersecurity
  • Finance
  • DeFi & Blockchain
  • Startups
  • Gaming
Dataconomy
  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Glossary
    • Whitepapers
  • Newsletter
  • + More
    • Conversations
    • Events
    • About
      • About
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
Subscribe
No Result
View All Result
Dataconomy
No Result
View All Result

How Your Tumblr Activity Gives Away Your Drinking Habits

byEileen McNulty
August 20, 2014
in Artificial Intelligence, News
Home News Artificial Intelligence

According to Gartner, 80% of data created today is unstructured. No data could be less structured than the vast ocean of selfies, photos of Starbucks beverages and millennial angst that is Tumblr. But the billion-dollar company may have found a way to take this rich, messy cultural cachet and turn it in to something profitable for big business- by using machine learning to mine its image content for brand sentiment analysis.

Tumblr have just inked a deal with Ditto, which will allow Ditto unfettered access to Tumblr’s content to scan for branded products, and sell the results to corporations like Coca-Cola, Jack Daniels- and, of course, the aforementioned Starbucks.

David Rose, CEO of Ditto, discussed the deal with Motherboard. “Twitter and Instagram have been suppliers of data to us for over a year and we’re able to look on behalf of brands at what the conversation about them looks like through photos,” he remarked. “What’s different about Tumblr is that they’re supplying us with the entire firehose of all photos that we’re then able to interpret.”

Their machine learning algorithm has been trained to recognise brands, with surprising accuracy; taking a look at their live stream will show you that even a corner of a logo, or a crumpled logo on a T-shirt, doesn’t go unrecognised. This information is then enriched with data about how influential a given user is. Although Rose is keen to stress: “We’re looking for the presences of faces, and the expressions of faces, but not for individual people.”

Brand-Affinities

Looking at the image above, the opportunity for brands is clear- they can see how far their reach is, which influential users are promoting their product, and even how their brand sentiment aligns with user’s tastes (Beyonce fans are evidently huge Smirnoff Ice drinkers, who knew?)

“Most people just don’t talk about most of the things in photos,” Rose said. “People never mention, like, ‘Me and a Corona at the beach.’ They just say, ‘The high point of my day,’ ‘Enjoying a great time,’ or, ‘Booyah.’ And so most of the signal is completely invisible to product companies.”

But what exactly is in it for Tumblr? Although the financial benefits of such a partnership are fairly evident, for a platform which facilitates a degree of anonymised self-expression, this could be viewed as a potential misstep. In the wake of growing discomfort about the way social media behemoths Facebook and OkCupid are handling user data, the partnership with Ditto could unsettle Tumblr’s sizeable community. As Motherboard notes, the deals perpetuates the impression that “users are being sold off to the highest bidder, without any kind of tangible return.” Of course the counterargument would be that these images are already being made publicly available, and compared to affecting your mood or meddling with your love life, Ditto’s work is comparatively benign.

Read more here.
Follow @DataconomyMedia

(Image credit: Ditto)

Stay Ahead of the Curve!

Don't miss out on the latest insights, trends, and analysis in the world of data, technology, and startups. Subscribe to our newsletter and get exclusive content delivered straight to your inbox.

Tags: Computer VisionMachine Learning NewsletterWeekly Newsletter

Related Posts

Is Grok 5 a revolution in AI or just Elon Musk’s latest overhyped vision?

Is Grok 5 a revolution in AI or just Elon Musk’s latest overhyped vision?

September 3, 2025
ICMP: Gemini, Claude and Llama 3 used music without any license

ICMP: Gemini, Claude and Llama 3 used music without any license

September 3, 2025
YouTube Premium cracks down on out-of-home family plans

YouTube Premium cracks down on out-of-home family plans

September 3, 2025
J-ENG unveils 7UEC50LSJA-HPSCR ammonia ship engine

J-ENG unveils 7UEC50LSJA-HPSCR ammonia ship engine

September 3, 2025
Judge rules Google won’t have to sell Chrome browser

Judge rules Google won’t have to sell Chrome browser

September 3, 2025
ShinyHunters uses vishing to breach Salesforce data

ShinyHunters uses vishing to breach Salesforce data

September 3, 2025
Please login to join discussion

LATEST NEWS

Is Grok 5 a revolution in AI or just Elon Musk’s latest overhyped vision?

ICMP: Gemini, Claude and Llama 3 used music without any license

YouTube Premium cracks down on out-of-home family plans

J-ENG unveils 7UEC50LSJA-HPSCR ammonia ship engine

Judge rules Google won’t have to sell Chrome browser

ShinyHunters uses vishing to breach Salesforce data

Dataconomy

COPYRIGHT © DATACONOMY MEDIA GMBH, ALL RIGHTS RESERVED.

  • About
  • Imprint
  • Contact
  • Legal & Privacy

Follow Us

  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Glossary
    • Whitepapers
  • Newsletter
  • + More
    • Conversations
    • Events
    • About
      • About
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
No Result
View All Result
Subscribe

This website uses cookies. By continuing to use this website you are giving consent to cookies being used. Visit our Privacy Policy.