Browser fingerprinting

Master thesis work for two students - 30 hec

Download PDF description.

Objective

The thesis work objective is to develop a statistical model capable of identifying a unique web browser between repeated sessions without storing any information at the client side such as tracking cookies.

Background

One essential metric when analyzing the performance of advertising campaigns is reach, or how many unique persons that were exposed to the advertising campaign. For different types of media (TV, print, online, etc.) there are different models to predict the reach of the campaign, each with its own limitations.

For online advertising, a common method used is to count the number of unique web browser cookies and assume these correspond to the number of unique persons. One major limitation to this approach is cookie deletion, private browsing and other user behaviors where several cookies are placed in the same browser, counting it multiple times.

The objective of this thesis work is to eliminate this limitation.

Preconditions - Browser fingerprinting

Each web browser provides certain compatibility information upon request. Examples of this information can be which version of each plug-in that is installed, the screen resolution of the computer or which Internet service provider that owns the IP-address.

Given enough detailed data about each browser, it will be possible to uniquely identify a vast majority of the web browsers with high confidence, without using cookies.

Task / Objectives

• Identify and validate non-intrusive data that can be collected and used for this purpose
• Based on real collected data, develop a model which can be used in near real-time to determine the uniqueness of a browser. The model will include definitions of which changes in the collected data indicate a new browser and which ones that likely indicate the same browser.
• Document the results in a written report in English

References

Panopticlick is a research project aimed at studying what types of information your browser gives away when prompted.

They published this paper on browser uniqueness: http://panopticlick.eff.org/browser-uniqueness.pdf

Details about the new ‘cookie law’ that was passed by EU and US during summer of 2011: http://bit.ly/nvqnLr

General information about ‘cookies’ used by all web sites: http://www.allaboutcookies.org/

Project Structure

The thesis work will be carried out at Burt AB, and should be started as soon as possible. The thesis is suitable for two students and consists of 30 hec credits.

Required knowledge / Competence

• High level programming, preferably Ruby and JavaScript
• Detailed technical knowledge of web browsers and Internet functionality
• Genuine interest in statistics, problem solving and online behavior

How to apply

Please send your resume / CV to: jobs@burtcorp.com

About Burt

We’re 24 people working around the clock to put a dent in the advertising industry by creating awesome tools for ad agencies, publishers and advertisers. Working for a startup is fun and fast paced. If you haven’t experienced it before, it’s like no job you’ve ever had.
If you’re constantly full of energy you’ll quickly find that this is a wonderful place to let all that energy out.

#Burt, #Master thesis, #Rich, #Rich Metrics, #browser fingerprinting, #hiring,