Guide to the Clickstream Data

1
Guide to the Clickstream Data

• Petr Berka
• University of Economics, Prague
• berka_at_vse.cz

2
Web Usage Mining Domain

• click-stream - a sequential series of page view
  (displays on users browser at one time)
  requests,
• server session - a click-stream of page views for
  a single user for a particular web site,
• user session - is the click-stream of page views
  for a single user across the entire web.

3
The Clickstream Data

• 3Millions of records (24 days) from a www shop
  web server log
• Contains information about time IP address
  session ID page request referer
• There are hundreds of thousands of sessions most
  of them very short, on average 16 pages
• Each page request in this www shop has the same
  structure page type / content ID (product ID)
• Page types are for example dp (detail of
  product), sb (shopping basket), ct (contact)

4
Example of the Data
unix time IP address session ID
page request referee 1074589200193.17
9.144.2 1993441e8a0a4d7a4407ed9554b64ed1/dp/?id
124 www.google.cz 1074589201194.213.35.234399
5b2c0599f1782e2b40582823b1c94/dp/?id182
1074589202194.138.39.56 2fd3213f2edaf82b27562d
28a2a747aa/ www.seznam.cz 1074589233
193.179.144.2 1993441e8a0a4d7a4407ed9554b64ed1/
dp/?id148 /dp/?id124 1074589245193.179.144.2
1993441e8a0a4d7a4407ed9554b64ed1/sb/
/dp/?id148 1074589248194.138.39.56
2fd3213f2edaf82b27562d28a2a747aa/contacts/
/ 1074589290193.179.144.2 1993441e8a0a4d7a4407e
d9554b64ed1/sb/ /sb/
5
Data Description

• table obchod (shop) - name of the internet shop
  (7 entries),
• table kategorie (category) - info about
  category of products (64 entries),
• table list (sheet) - info about a specific
  product of a more detailed type (157 entries),
• table znacka (brand) - name of the producer or
  brand of a product (197 entries),
• table tema (theme) - info about themes
  discussed in the on-line advice (36 entries)

6
Data Summary (1/3)

• 3 617 171 page requests

• 522 410 sessions
• 318 523 single page
• 203 887 length gt 1
• avg. length 16
• median 8
• modus 2
• longest 15454

7
Data Summary (2/3)

• time spent during a session
• avg. time 002446
• median 000308
• modus 000009
• longest 4332753

8
Data Summary (3/3)
distribution of sessions with length gt 1