The social network will start collecting data from the first minute the user interacts with its content. For example, this new technology will give Facebook the ability to see where the user's cursor moves or whether its newsfeed is visible at a given moment on the screen or on the user's mobile screen, Ken Rudin reports on Tuesday during his interview.
Mr. Rudin added that further deepening of information gathering will help in product development and in more precise targeting of an advertising campaign.
Facebook collects two types of data: demographic and behavioral. Demographic data, such as where the user lives or which school they attended, i.e., user information beyond what is directly related to the internet. The “data for behavior» are, for example, the circle of friends on Facebook, the likes a person makes and how these are reflected in real time on the network. Ongoing tests will greatly expand the data collected about user behavior according to Mr. Rubin. The tests are ongoing and are part of a broader technological program.
These new ways of data collection will make Facebook know if the mouse cursor passed over a specific ad, said Rubin, a former Silicon Valley executive who joined Facebook in April 2012 from Zynga INC where he had served as vice president at the analytics and platform technologies level
Facebook is preparing to change the company's infrastructure so it can increase its data volume. Facebook is not the first company to develop such activity. Shutterstock Inc, a marketplace for digital images, literally records everything users do. This company uses the open-source Hadoop, a system for data analysis that can show how far the mouse cursor is over an image.
Facebook has increased 4,000 times over the past four years the data it stores. The data has reached the level of 300 petabytes. The company uses the HADOOP file distribution system we mentioned earlier to manage its data according to Rudin.
The most important thing for this data warehouse is that it be accessible to companies and that someone can easily choose what they want. If this is not possible, the data are useless. Rubin states that for this purpose there is a plan to create an index, essentially a search engine for this data.
By Steve Rosenbush
Translation – adapted from the wall street journal – Cleo Ioannou
Source: koutipandoras.gr
