Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Replicating Data Mining Techniques for Development: A Case Study of Corruption

Ransom, James LU (2013) MIDM71 20131
LUMID International Master programme in applied International Development and Management
Abstract
Data Mining has a reputation in social science for lacking statistical rigour. This study challenges this reputation and argues that, whilst such a method (as with any other) can be abused, it has particular promise as a tool to be used for monitoring and explorative research, especially by smaller development organisations. Drawing on recent advances in adapting commercial ‘Big Data’ techniques for use in international development, this study uses an example data set of global news reports to measure the level of discussion about corruption using a Text Mining methodology. The methodology outlined holds particular promise for tracking the dissemination of ideas and concepts, although it is heavily dependent on contextual interpretation... (More)
Data Mining has a reputation in social science for lacking statistical rigour. This study challenges this reputation and argues that, whilst such a method (as with any other) can be abused, it has particular promise as a tool to be used for monitoring and explorative research, especially by smaller development organisations. Drawing on recent advances in adapting commercial ‘Big Data’ techniques for use in international development, this study uses an example data set of global news reports to measure the level of discussion about corruption using a Text Mining methodology. The methodology outlined holds particular promise for tracking the dissemination of ideas and concepts, although it is heavily dependent on contextual interpretation and the quality of the data set used. (Less)
Please use this url to cite or link to this publication:
author
Ransom, James LU
supervisor
organization
course
MIDM71 20131
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Corruption, Big Data, Data Mining, Bribery, Text Mining, Civil Society Organisations, Broadcast Media, Data Analysis
language
English
id
3798253
date added to LUP
2013-06-27 11:18:22
date last changed
2013-06-27 11:18:22
@misc{3798253,
  abstract     = {Data Mining has a reputation in social science for lacking statistical rigour. This study challenges this reputation and argues that, whilst such a method (as with any other) can be abused, it has particular promise as a tool to be used for monitoring and explorative research, especially by smaller development organisations. Drawing on recent advances in adapting commercial ‘Big Data’ techniques for use in international development, this study uses an example data set of global news reports to measure the level of discussion about corruption using a Text Mining methodology. The methodology outlined holds particular promise for tracking the dissemination of ideas and concepts, although it is heavily dependent on contextual interpretation and the quality of the data set used.},
  author       = {Ransom, James},
  keyword      = {Corruption,Big Data,Data Mining,Bribery,Text Mining,Civil Society Organisations,Broadcast Media,Data Analysis},
  language     = {eng},
  note         = {Student Paper},
  title        = {Replicating Data Mining Techniques for Development: A Case Study of Corruption},
  year         = {2013},
}