keimhaqi / tomorrow

Fingerprinting anonymous users and estimating unique visitors, using a HyperLogLog

Tomorrow

Close your eyes and I'll kiss you, tomorrow I'll miss you -- The Beatles

Experimental service to provide a count of unique ANONYMOUS visitors to a website.

Not having access to a unique id (cookie, session, etc), we attempt to fingerprint a user through examining their browser topology (device fingerprinting -- see https://panopticlick.eff.org/)

Now that we have a lot of fingerprints, how do we efficiently store them, and get unique counts? For this we use a HyperLogLog, as implemented in Redis.

require('tomorrow').start({
	port : 8081, // Optional. Port to run server on.
	redisPass : "foobar", // Optional. Redis password.
	redisPort : 6379 // Optional. Port Redis is running on.
})

This starts a server with three routes:

• /test - will load test.html file so you can play with library
• /fingerprint - returns a js library that takes user fingerprints.
• /log/:fingerprint - sending a :fingerprint hash to /log stores the print.
• /count - returns the current cardinality (estimate of unique fingerprints)

On the client, get the library via /fingerprint, then run this JavaScript command to get the current client's unique fingerprint:

new Fingerprint().get()

Optionally, you can add more accuracy with canvas:

new Fingerprint({canvas: true}).get())

( http://cseweb.ucsd.edu/~hovav/dist/canvas.pdf )