Predictive Customer Data Analytics : The Value of Public Statistical Data and the Geographic Model Transferability
Preprint
- 1 January 2018
- preprint
- Published by Universitatsbibliothek Bamberg
Abstract
Companies pay high prices for detailed customer information (e.g., income, household type) for gaining insights and conducting targeted marketing campaigns. We argue that companies can utilize predictive analytics artifacts to derive such information from existing customer data in combination with freely available data sources, such as open government data. In this study, we use a machine learning artifact for a specific yet highly relevant case from the utility industry, trained on data of 7,504 energy customers and investigate two important aspects for predictive business analytics: First, we identified the sparsely available open government statistics and found that even that limited amount of open data can increase our artifact's performance. Second, we applied the predictive models, trained with a regional customer dataset, on households in other geographic regions with acceptable performance loss. The results support the development of systems aiding managerial decision-making, predictive marketing and showcase the value of open data.Keywords
This publication has 29 references indexed in Scilit:
- Revealing household characteristics from smart meter dataEnergy, 2014
- Feature extraction and filtering for household classification based on smart electricity meter dataComputer Science - Research and Development, 2014
- Transforming decision-making processes: a research agenda for understanding the impact of business analytics on organisationsEuropean Journal of Information Systems, 2014
- Spatial transferability of travel forecasting models: a review and synthesisInternational Journal of Advances in Engineering Sciences and Applied Mathematics, 2013
- ‘Datafication’: making sense of (big) data in a complex worldEuropean Journal of Information Systems, 2013
- The impact of consumers’ feedback preferences on domestic electricity consumptionApplied Energy, 2012
- Transferability of accident prediction modelsSafety Science, 2006
- Comparing two K-category assignments by a K-category correlation coefficientComputational Biology and Chemistry, 2004
- The theory and practice of decoupling utility revenues from salesUtilities Policy, 1997
- Testing Transferability of Recreation Demand Models Across Regions: A Study of Corps of Engineer ReservoirsWater Resources Research, 1995