Data Cleaning Methods for Client and Proxy Logs

H. Weinreich; H. Obendorf; E. Herder

Data Cleaning Methods for Client and Proxy Logs

H. Weinreich, H. Obendorf, E. Herder

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

38 Downloads (Pure)

Abstract

In this paper we present our experiences with the cleaning of Web client and proxy usage logs, based on a long-term browsing study with 25 participants. A detailed clickstream log, recorded using a Web intermediary, was combined with a second log of user interface actions, which was captured by a modified Firefox browser for a subset of the participants. The consolidated data from both records revealed many page requests that were not directly related to user actions. For participants who had no ad-filtering system installed, these artifacts made up one third of all transferred Web pages. Three major reasons could be identified: HTML Frames and iFrames, advertisements, and automatic page reloads. The experiences made during the data cleaning process might help other researchers to choose adequate filtering methods for their data.

Original language	Undefined
Title of host publication	Workshop on Logging Traces of Web Activity: The Mechanics of Data Collection
Editors	A. Edmonds, K. Hawkey, M. Kellar, D. Turnbull
Place of Publication	Halifax, Canada
Publisher	Dalhousie University
Pages	-
Number of pages	4
ISBN (Print)	not assigned
Publication status	Published - 23 May 2006

Publication series

Name
Publisher	Dalhousie University
Number	We.C1.3

Keywords

EWI-9155
HMI-IE: Information Engineering
IR-66893
METIS-237939
HMI-HF: Human Factors

Access to Document

http://torch.cs.dal.ca/~www2006/programme.htm

Cite this

@inproceedings{0f842e2874604915bbfae7ed5f944e66,

title = "Data Cleaning Methods for Client and Proxy Logs",

abstract = "In this paper we present our experiences with the cleaning of Web client and proxy usage logs, based on a long-term browsing study with 25 participants. A detailed clickstream log, recorded using a Web intermediary, was combined with a second log of user interface actions, which was captured by a modified Firefox browser for a subset of the participants. The consolidated data from both records revealed many page requests that were not directly related to user actions. For participants who had no ad-filtering system installed, these artifacts made up one third of all transferred Web pages. Three major reasons could be identified: HTML Frames and iFrames, advertisements, and automatic page reloads. The experiences made during the data cleaning process might help other researchers to choose adequate filtering methods for their data.",

keywords = "EWI-9155, HMI-IE: Information Engineering, IR-66893, METIS-237939, HMI-HF: Human Factors",

author = "H. Weinreich and H. Obendorf and E. Herder",

year = "2006",

month = may,

day = "23",

language = "Undefined",

isbn = "not assigned",

publisher = "Dalhousie University",

number = "We.C1.3",

pages = "--",

editor = "A. Edmonds and K. Hawkey and M. Kellar and D. Turnbull",

booktitle = "Workshop on Logging Traces of Web Activity: The Mechanics of Data Collection",

}

Data Cleaning Methods for Client and Proxy Logs. / Weinreich, H.; Obendorf, H.; Herder, E.

Workshop on Logging Traces of Web Activity: The Mechanics of Data Collection. ed. / A. Edmonds; K. Hawkey; M. Kellar; D. Turnbull. Halifax, Canada : Dalhousie University, 2006. p. -.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Data Cleaning Methods for Client and Proxy Logs

AU - Weinreich, H.

AU - Obendorf, H.

AU - Herder, E.

PY - 2006/5/23

Y1 - 2006/5/23

N2 - In this paper we present our experiences with the cleaning of Web client and proxy usage logs, based on a long-term browsing study with 25 participants. A detailed clickstream log, recorded using a Web intermediary, was combined with a second log of user interface actions, which was captured by a modified Firefox browser for a subset of the participants. The consolidated data from both records revealed many page requests that were not directly related to user actions. For participants who had no ad-filtering system installed, these artifacts made up one third of all transferred Web pages. Three major reasons could be identified: HTML Frames and iFrames, advertisements, and automatic page reloads. The experiences made during the data cleaning process might help other researchers to choose adequate filtering methods for their data.

AB - In this paper we present our experiences with the cleaning of Web client and proxy usage logs, based on a long-term browsing study with 25 participants. A detailed clickstream log, recorded using a Web intermediary, was combined with a second log of user interface actions, which was captured by a modified Firefox browser for a subset of the participants. The consolidated data from both records revealed many page requests that were not directly related to user actions. For participants who had no ad-filtering system installed, these artifacts made up one third of all transferred Web pages. Three major reasons could be identified: HTML Frames and iFrames, advertisements, and automatic page reloads. The experiences made during the data cleaning process might help other researchers to choose adequate filtering methods for their data.

KW - EWI-9155

KW - HMI-IE: Information Engineering

KW - IR-66893

KW - METIS-237939

KW - HMI-HF: Human Factors

M3 - Conference contribution

SN - not assigned

SP - -

BT - Workshop on Logging Traces of Web Activity: The Mechanics of Data Collection

A2 - Edmonds, A.

A2 - Hawkey, K.

A2 - Kellar, M.

A2 - Turnbull, D.

PB - Dalhousie University

CY - Halifax, Canada

ER -