Towards big-data analysis of deviation and error reports in product development projects
DS 85-2: Proceedings of NordDesign 2016, Volume 2, Trondheim, Norway, 10th - 12th August 2016
Year: 2016
Editor: Boks, Casper; Sigurjonsson, Johannes; Steinert, Martin; Vis, Carlijn; Wulvik, Andreas
Author: Arnarsson, Ivar Örn; Malmqvist, Johan; Gustavsson, Emil; Jirstrand, Mats
Series: NordDESIGN
Institution: 1: Chalmers University of Technology, Sweden; 2: 2Fraunhofer Chalmers Centre (FCC) for Industrial Mathematics, Sweden
Section: Methodology: General Applications
Page(s): 083-092
ISBN: 978-1-904670-80-3
Abstract
Background
In large complex system development projects late changes are unfortunately a common thing. It is well known that changes made later in a development project are very costly and cause delays (Clark and Fujimoto 1991). The root causes of these changes are still poorly understood. Though the severity of the problem gained attention already in the 1980’s, a recent study by Giffin et al. (2009) confirms that this problem still poses challenges.
Aims & research questions
The aim of the paper is to develop big-data based tools to analyze product deviations and to connect them to their originating root causes, enabling the root causes to be eliminated rather than the symptoms.
We address two research questions:
1. Is it possible to use product deviations reports from databases and apply “big data” analysis to find patterns that will aid in finding root causes for deviations and repeatable failures?
2. Is it possible from this analysis to build fact-based hypotheses for how to improve product quality in the late phases of a project?
Approach/Methods used
Big data text analysis tools will be employed to analyse and find patterns in product deviations in late phases of a product development project. Deviations will be classified with respect to type, truck location, time, frequency, severity, change requests etc. These patterns will bring facts on the table that are not visible looking at the raw data and thereby make it possible to correct repeated failures. The data sources are from a large, recently concluded, truck development project. They include a prototype build and test report database, a product documentation database, and steering committee meeting protocols.
Preliminary conclusions/Expected findings
The data analysis will provide the basis for building a set of hypotheses about root causes and late changes. A broad, corporate-level, understanding of late changes will be obtained. This type of analysis has not been performed at the truck manufacturer before, and, as far as we know not elsewhere either.
Keywords: product development project, big-data analysis, root causes