Areas of research: social network topography, online activism, information and communications technologies, international relations, influence qunatification, telemetry
Practical Experience: API hackery, distributed computing, semantic web technologies, database awesomeness
Senior Developer/R&D guy, Little Bird, Portland, OR — 2012-current
To say that Marshall Kirkpatrick is respected in the Internet startup ecosystem would be a significant disservice to him – he has been hugely involved in all things tech for the better part of a decade (at least), and is probably one of the single most well known, well connected, and well positioned people to run a startup. When I heard that he was doing this, I was chomping at the bit. As the first hire at Little Bird, my job is manifold – mostly, I try to spend as little time as possible on the boring bits – I’m constantly pushing us into territory that I haven’t seen other startups go in, and try to quickly identify if there’s value in the direction I’m pushing (and abandon it if its not going to be efficacious). It’s fun, super hard work, fast paced, but also the freest ecosystem I’ve ever seen for a company. Come get at our stuff here.
Consultant, SocialFlow, New York, NY — 2011
I have been picked up to do a series of research/analytics based projects for a company known as SocialFlow, based out of New York. Gilad Lotan and I basically sit in front of mounds of tweets and figure out how to better leverage the data for SocialFlow’s needs, and on the side we do really fun, useful, nearly academic-level (we get lazy in the fourth quarter) projects.
Developer (Wizard), 140Proof, San Francisco, CA — 2011
I’m building lots of cool stuff for a targeted ad platform for Twitter. Basically, I sit down and get to think up how best to match content to users, then go off and formally implement that thought. If it works better than the way things were working before, its a success and I start over again. If its not, its a failure and I start over again. I also do a bunch of maintenance and general development for the platform with Jamie Wilkinson, who I’ve known for years now and am very fond of working with.
Intern, Berkman Center for Internet and Society, Harvard Law School, Cambridge, MA — 2010
I was tasked with two things at the Berkman Center – build a relatively straightforward Rails app for use as an education tool (teach kids about fair use and copyright), and collecting Twitter data. I can’t talk much about the actual data collected, as the paper is still in process, but I can say this: we were able to grab the bulk of Tweets from a specific country, over the entire existence of Twitter, and do analytics on this data set, all without any special access from Twitter. We were able to play disasters in reverse, and see cool correlations with data that have never had a parallel, as far as I’ve seen. Hopefully they find the right place to publish this information someday.
Managing Director, Web Ecology Project, Cambridge, MA — 2010
With his departure from Cambridge and to California, Tim Hwang put Sam Gilbert and I in charge of Web Ecology in his stead. Currently, the Web Ecology Project works on a few fronts – generating interesting research papers not necessarily under the jurisdiction of traditional academics, creating useful programs, applications, and frameworks for better research into online phenomena, and meeting face to face for rapid prototyping of projects in our quarterly meetings, traditionally either in Boston or New York.
Visiting Researcher, Tetherless World Constellation, Rensselaer Polytechnic Institute, Troy, NY — 2010
Create demonstrations showcasing semantic web technologies and employing predominantly US Government data sets available at http://data.gov. Created visualization templates in Processing, Flex, and adapted templates using the Google Visualization API. Additionally, created tutorials and walkthroughs on how any user could adapt SPARQL queries and templates to create new visualizations as part of an outreach component of the work done at TWC.
Intern, Rocketboom Inc; New York, NY — 2009
Responsible for development on http://KnowYourMeme.com, developing screen scraping and data collection library for http://mag.ma/. Collaborated with Jamie Wilkinson and others on developing appropriate back-end for http://mag.ma, which is an online video aggregator site, collecting information about online videos, then ostensibly placing them in a chart system to determine popularity.
Interactive Developer, Instrument Marketing; Portland, OR — 2007-2010
A wide range of duties while working at Instrument Marketing have been: Content Management, development of CSV parsers for proprietary CMS software, implementation of various reservation systems. Daily work included updating/templating/managing various client websites (viewable at http://www.weareinstrument.com) as well as creating in-house tools for CMS.
Oxford University, Oxford, UK — MSc in Social Science of the Internet, 2012
Title:“Determining the Importance of Geography on Twitter”
Abstract: Information and Communications technologies have always been held as agents capable of vastly shifting the geographic distribution of personal communication networks for various uses – in the Internet age, the claims are even more exceptional. While the Internet theoretically allows users to communicate without geographic considerations, what are the practical contours of this geographical distribution? This question is investigated through an in-depth analysis of one domain of use, Twitter, where many communication ties are explored via their geographic, social, and semantic components tying agents together in an effort to understand what factors augment traditional geographically distributed communication.
Bennington College, Bennington, VT — BA, 2010
Title:”#iranelection: Are the dynamics and structure of Web 2.0 shifting global policy in 140 characters or less?”
Abstract: Using a data set of 766,263 tweets captured between June 12th, 2009, and October 25th, 2009, matching the Twitter category (or “hashtag”) of #iranElection, what new insights can be gained? In previous studies surrounding online activism, qualitative, systemic, or at-a-distance analysis tends to predominate. In a Web 2.0 environment, and with new programming frameworks that stress rapid prototyping, how can we gain novel insights into the nature of online activism, citizen journalism, and the role of the internet in our society and politics at large?
(2012) Statistical Probability That Mitt Romney’s New Twitter Followers Are Just Normal Users: 0%. Furnas, A. and Gaffney, D. In: The Atlantic. July 31st, 2012.
(2013) Data collection on Twitter Puschmann, C., & Gaffney, D. In K. Weller, A. Bruns, J. Burgess, M. Mahrt & C. Puschmann (eds.), Twitter and Society. New York, NY: Peter Lang.
(2014) The Multiple Facets of Influence. Dubois, E. and Gaffney, D. American Behavioral Scientist 0002764214527088, first published on April 8, 2014 doi:10.1177/0002764214527088
(2013) Identifying the opinion leader: Influence, Twitter, and Canadian politics. Dubois, E. and Gaffney, D.. #Influence13 Symposium & Workshop. Dalhousie University, Halifax, Nova Scotia.
(2012) Game or measurement? Algorithmic transparency and the Klout score. Gaffney, D. and Puschmann, C. #Influence12 Symposium & Workshop. Dalhousie University, Halifax, Nova Scotia. (Video)
(2012) Where in the world are you? Geolocation and language identiﬁcation in Twitter Hale, Scott, Devin Gaffney, and Mark Graham. Forthcoming.
(2011) The Revolutions Were Tweeted: Information Flows during the 2011 Tunisian and Egyptian Revolutions Lotan, Gilad, Erhardt Graeff, Mike Ananny, Devin Gaffney, Ian Pearce and danah boyd. International Journal of Communcation Web. 8 Sep. 2011.
(2010) #iranElection: quantifying online activism. In: Proceedings of the WebSci10: Extending the Frontiers of Society On-Line, April 26-27th, 2010, Raleigh, NC: US.
(2010) ChatRoulette: An Initial Survey Web Ecology Project, March 1st, 2010
(2010) The data your data could smell like Web Ecology Project, September 27th, 2010
(2013) Location! The Importance of Geo-data, with Mark Graham, Monica Stephens, and Catherine D’Ignazio. , Presented at South by Southwest, March 12th, 2013, Sheraton Hotel, Austin, Texas.
(2012) Ditching the dowsing rod – Why prescribed analytics fail to deliver (Video), Presented at the Eduserv Symposium 2012, May 10th, 2012, Royal College of Physicians, London
140kit – A distributed computing network for recording live Twitter Search API data. A user inputs a term, a length of time to record data for that term, and they are sent to a URL that allows Researchers to request analytical processes to be conducted against the data, the flexibility to define their own options for what will be conducted on the dataset, the ability to review individual tweets within the dataset, and a rich environment to generally allow researchers to quickly understand the dimensions of a phenomenon on Twitter.
GPS Balloon – A physical computing project which consisted of an Arduino project board controlling a cell phone, taking readings from a GPS sensor, reading temperature sensor data, and committing this all to SD card memory. A camera was attached to the bottom of the project (and was controlled by the Arduino) so that aerial images could be retrieved upon collection. The cell phone would then send a message to Twitter, in order to transmit coordinates so that the project could be retrieved and SD card read. The total cost for this project was < $200, and once retrieved, the project could be redeployed immediately.
Wikipedia Network Maps – This is a program that scrapes data from Wikipedia in order to create visual maps of linkages between articles. Essentially, the program seeds a random article, then recurses through an entire connected component. When it runs out of nodes, it seeds another random article, and continues this process until it has mapped out an entire Wikipedia network. It was never tested for scalability to large Wikipedia networks such as the English/Japanese/French/German/Spanish subdomains, but does accurately plot Wikipedia of article number n < 5,000. In general, the idea was to compare 14 small Wikipedia networks to draw conclusions on how users organize data online.
Proficient: C, Java
Conversational: Actionscript, Assembly, C#, Objective C
Protocols: TCP/IP (can build implementations), SMTP (general work)
Markup/Object Notation/Templating languages: HTML, XML, JSON, CSS, Textile, RDF
Query Languages: SPARQL, SQL, MongoDB
Understanding of RPC systems, Hadoop, Amazon EC2 and S3, MapReduce algorithm
(2012) Ehrenberg, Rachel. “Social Media sway”, October 20th, 2012
(2012) Condliffe, Jamie. “Evidence That Klout Doesn’t Make Sense, Visualized”, Gizmodo.com, July 18th, 2012
(2012) Tortello, Letizia. “A Start up weekend vince Burst l’invenzione che cambierà le news”, La Stampa, June 11th 2012
(2011) Watters, Audrey. “How Recent Changes to Twitter’s Terms of Service Might Hurt Academic Research.”, Read Write Web March 3rd, 2011.
(2011) Kolowich, Steve. “Twitter Fortifies Nest”, Inside Higher Ed. March 8th, 2011 (Re-syndicated at New York Times, American Public Media, UMBC Ebiquity)
(2011) Taylor, Chris. “How Bin Laden News Exploded on Twitter: A Visualization”, Mashable May 6th, 2011.
(2010) “New Scientist: Exploring the Network without the Guesswork”, May 10th, 2010, Issue 2759
Also I just learned my Erdos number is 5. That’s neat.