Promoted tweets versus promoted trends

Advertising on Twitter? There are a number of factors to consider: use of hashtags, use of keywords, target audience and locations. But also choosing the type of adverts is important: promoted tweets, or promoted trends? Our research shows that promoted trends lead to higher number of tweets, while promoted tweets lead to higher user engagement, which may be the ultimate measure of success. For a use-case study, see:


Shana Dacres, Hamed Haddadi, Matthew Purver, “Topic and Sentiment Analysis on OSNs: a Case Study of Advertising Strategies on Twitter”, arXiv, December 2013 (PDF)


Meeyoung Cha, Hamed Haddadi, Fabricio Benevenuto, Krishna Gummadi, "Measuring User Influence in Twitter: The Million Follower Fallacy", in ICWSM 2010, 4th Int'l AAAI Conference on Weblogs and Social Media, May 23-26, 2010, George Washington University, Washington, DC (paper) (NYTimes coverage ) (Harvard Business Review coverage ) (Media coverage)

Human-Data Interaction

I have been thinking about this concept a lot recently... just to clarify, this is NOT a new discipline or field, it is basically an interest group, and a way of thinking about the aggregate effect of user profiling, advertising, interface, HCI, privacy and the way the users' behaviour may be changed due to change in these topics, rather similar to the way that the Quantified Self movement is shaping up.. an interest group of inter-disicplinary researchers..

This all started after a casual talk and meeting with Mort in Nottingham, then filtering the ideas through Jon, Mac and Tristan, which resulted to a draft of our thoughts for the UK Digital Economy all-hands meeting:

 Richard Mortier, Hamed Haddadi, Tristan Henderson, Derek McAuley, and Jon Crowcroft, "Challenges & Opportunities in Human-Data Interaction", DE2013: Open Digital, November 2013, MediaCityUK, Salford, UK. (PDF)


On 4th October 2013, we had the EPSRC/ITaaU first workshop on HDI, which brought in a large number of top-notch industry and academic researchers across law, privacy, HCI, economics, computer science, engineering, geography, and financial sector. You may enjoy reading the detailed liveblog by Cambridge guys. The participants all took part in a range of activities, and are interested in officially forming an interest group, so watch this space for more news!

Individual's perception of the value of privacy, a contextual experiment.

There are a large number of studies surveying individuals about the value fo their personal information. These are particularly motivated by the claims of big cloud , good or bad, that individuals can't evaluate their personal information, so it must be free!

we recently did a survey using android apps , asking individuals to let us know twice a day, what they are doing, who they are with, and how much are these information worth to them. However we divided the individuals into 4 groups, with different buying and selling criteria..

In brief, the results show that individual's CAN evaluate the value of their personal info, and these values are in agreement with findings of others (see Vijay Erramilli's paper in WWW'13 for example or Bernardo Huberman's information market paper)

The paper is published at ACM SIGCOMM HOTPLANET 2013 workshop. you can ready it here

Bernadette Kamleitner, Stephan Dickert, Marjan Falahrastegar, Hamed Haddadi, “Information Bazaar: a Contextual Evaluation”, 5th ACM HotPlanet workshop, co-located with SIGCOMM 2013, August 2013, HongKong.