Abstract
Continuous delivery is an activity in the field of continuous software engineering. Data warehousing, on the other hand, lie within information systems research. This dissertation combines these two traditionally separate concerns of continuous delivery and data warehousing.
This dissertation’s motivation stems from a practical problem: how to shorten the time from a reporting idea until it is available for users. Data warehousing has traditionally been considered tedious and delicate. In data warehousing, distinct steps take place one after another in a predefined unalterable sequence. Another traditional aspect of data warehousing is bringing everything at once to a production environment, where all the pieces of a data warehouse are in place before production use. If development follows agile iterations, why are the releases in production not following the same iterations?
This dissertation introduces how reporting and data warehouse teams can synchronously build business intelligence solutions in increments. Joint working enhances communication between developers and shortens the feedback cycle from an end-user to developers, and makes the feedback more direct. Continuous delivery practices support releasing frequently to a production environment. A two-layer data warehouse architecture separates analytical and transactional processing. Separating different processing targets enables better testing and, thus, continuous delivery. When frequently deploying with continuous delivery practices, automating transformation creation in data warehousing reduces the development time. This dissertation introduces an information model for automating the implementation of transformations, getting data into a data warehouse and getting data out of it.
The research evaluation followed the design science guidelines. Research for this dissertation collaborated with the industry. These ideas have been tested on real projects with promising results, and thus they have been proven to work.
This dissertation’s motivation stems from a practical problem: how to shorten the time from a reporting idea until it is available for users. Data warehousing has traditionally been considered tedious and delicate. In data warehousing, distinct steps take place one after another in a predefined unalterable sequence. Another traditional aspect of data warehousing is bringing everything at once to a production environment, where all the pieces of a data warehouse are in place before production use. If development follows agile iterations, why are the releases in production not following the same iterations?
This dissertation introduces how reporting and data warehouse teams can synchronously build business intelligence solutions in increments. Joint working enhances communication between developers and shortens the feedback cycle from an end-user to developers, and makes the feedback more direct. Continuous delivery practices support releasing frequently to a production environment. A two-layer data warehouse architecture separates analytical and transactional processing. Separating different processing targets enables better testing and, thus, continuous delivery. When frequently deploying with continuous delivery practices, automating transformation creation in data warehousing reduces the development time. This dissertation introduces an information model for automating the implementation of transformations, getting data into a data warehouse and getting data out of it.
The research evaluation followed the design science guidelines. Research for this dissertation collaborated with the industry. These ideas have been tested on real projects with promising results, and thus they have been proven to work.
| Original language | English |
|---|---|
| Place of Publication | Tampere |
| Publisher | Tampere University |
| ISBN (Electronic) | 978-952-03-2653-1 |
| ISBN (Print) | 978-952-03-2652-4 |
| Publication status | Published - 2022 |
| Publication type | G5 Doctoral dissertation (articles) |
Publication series
| Name | Tampere University Dissertations - Tampereen yliopiston väitöskirjat |
|---|---|
| Volume | 706 |
| ISSN (Print) | 2489-9860 |
| ISSN (Electronic) | 2490-0028 |