Open AccessHow to succeed in data science projects industrialization ?
Author(s) -
Hervé Potelle,
Laurent Leblond
Publication year - 2018
Publication title -
management and data science
Language(s) - English
Resource type - Journals
ISSN - 2555-7033
DOI - 10.36863/mds.a.4717
Subject(s) - mandate , data quality , process management , weighting , transformation (genetics) , computer science , business , engineering management , knowledge management , engineering , operations management , political science , medicine , metric (unit) , law , radiology , biochemistry , chemistry , gene
Industrializing Data Science projects in business lines results from a transformation that takes place from strategic scoping to operational management. By using a learning base, it is possible to predict with performance the success or the failure of future projects in the pre-scoping phase. Industrializable projects can thus be correctly predicted by weighting the following six criteria: the business question, the business mandate, the business availability, the "Data" competences, the quality / Data quantity and the Data monitoring. On this core, the design of a Predictive Score Card evaluation tool allows an optimized projects pre-scoping.