| Title | Mining Consumer-Generated Product-Configuration Data |
| Authors | Fuhrmann, Tino, Schweizer, Marvin, Geyer-Schulz, Andreas and Kurz, Peter |
| Year | 2024 |
| Volume | 8(1) |
| Abstract | Many companies offer their customers the possibility to configure their products online and thus to tailor them to their needs and wants. Today, this capability is implemented via sophisticated product-configurators which in parallel record those consumer-generated configurations. However, while a huge volume of such configuration data is recorded, little research has been conducted in the field of mining such configuration data. This is at least in part due to the non-availability of open configuration data sets. For the workshop of the special interest group for data analysis of the German Classification Society in Karlsruhe on 20–21 November 2015, the German market research company TNS Infratest has made available an anonymized configuration data set of a major car manufacturer. In this contribution a first report on mining this data set is given. We show that the development of data models which are specific for the analysis planned reduce the resource requirements several orders of magnitude, both for data storage and processing time. Furthermore, we use ordinal utility theory in iso-price segments of cars to operationalize the detection of interesting patterns: We detect irrational car configurations of consumers and the attribute sets for a rational choice. We show and discuss the use of such attribute combinations for sales communication in the purchase process of cars. |