Agil testning i datalagringsprojekt

Testing in data warehousing has traditionally been done in large chunks at the end of development projects. This has often led to costly bug fixes because fixing bugs that are found late in the development process is generally a very time consuming task. In some cases, there is not enough time to fix these bugs because of narrow project deadlines. This can have major consequences for organizations because business decisions often relies on their data warehouse to provide accurate statistics and data. This study examines a department of a major Swedish agency that recently adopted agile methods to overcome some of the shortcomings associated with traditional data warehouse development. The effects of agile methods on data warehouse testing are examined, as well as the challenges that are needed to overcome for them to be successful. We also touch upon consequences of poor test design and what implications incorrect data that are delivered from a data warehouse can have both internally and externally. In conclusion, the study have found that adopting agile methods have led to test activities being carried out more often, more efficiently and with better results than in the past. In addition, product owners have been continuously involved in the development process through frequent acceptance tests. For agile methods to be successful in data warehouse testing, we also concluded that they need the support of tools, system environments and attitudes both among developers and product owners. ; Nej