.conf Go 2023 - Data analysis as a routine
•
0 likes•290 views
.conf Go 2023 presentation: Data analysis as a routine Speaker: Markus Gustafsson - Senior Development Lead, Posti Group
Report
Share
More Related Content
Similar to .conf Go 2023 - Data analysis as a routine
Similar to .conf Go 2023 - Data analysis as a routine (20)
More from Splunk
More from Splunk (20)
Recently uploaded
Recently uploaded (20)
.conf Go 2023 - Data analysis as a routine
- 1. Data analysis as a routine Co-evolution of information system and data analysis development
- 2. INTERNAL • Posti Group • Posti was established in Finland in 1638 by Per Brahe. • Delivery and logistics company mainly in Finland, Sweden and Baltics • Almost 20.000 employees • Visiting 3M house holds and organizations in Finland every weekday. • A huge decrease in mail volumes due to the digitalization in our society and worldwide. • Efficiency improvements are sought through increased automation, supported by data analytics. • Me • Working in Posti since 1996 with information systems, mostly in development tasks. • Experience in different business groups, services, and customers including both consumers and commercial / non-commercial organizations. • Education: • M.Sc., major in Computer Science and first minor in Psychology • Doing post graduate studies about analytics as a routine in information systems development. About Posti and me
- 3. INTERNAL How to continuously improve the matching in our mail sorting process? A concept of the co-evolution regarding the development of Information System and data analysis Topics of the presentation 13.9.2023 3
- 4. INTERNAL 13.9.2023 4 Getting my Donald Duck The view from the subscriber Sending my Donald Duck The Donald Duck ”company” Me Reading my Donald Duck with a sandwich and juice Marcus Gustafsson
- 5. INTERNAL 13.9.2023 5 Getting my Donald Duck What really happens Sender, for example, a magazine publisher Printing house or mailer Advance information of recipients Machine sorting Delivery Customer service Receicing mail Printing and preliminary sorting Trasportation of mail items to Posti Inspection of incoming items Manual sorting Sender & mailer / printing house Posti Mail recipient Marcus Gustafsson To whom and to what address is this mail to be delivered? Sorting Delivery
- 6. INTERNAL 13.9.2023 6 To whom and to what address? Matching gives the answer the fundamental capability in sorting Matti Meikäläinen Tuohikatu 49 C 20 28100 PORI Matching 718910 ~1 000 million Matti Meikäläinen custno: 290189 c/o Sanni Meikäläinen PL 192, Tuohikatu 49 C 20 28100 PORI Sorting 2371891
- 7. INTERNAL Questions we are ask almost daily: • How good are we? • What problems (potential targets for improvements) do we have? • What is the expected benefit of solving a problem? • What part of the problem can be solved in a cost-efficient way (80/20)? • What do our customers prefer? 13.9.2023 7 How to improve matching? Answers to questions can be found by analyzing the production data continuously
- 8. INTERNAL 13.9.2023 8 How do we do analytics? Inspection Sorting production Systems development Printing house co-operation Master data maintenance Sales Legal Address clearance “Breaking” silos Virtual team The strength of different expertise, personalities and world views and at the same time hampers the creation of shared understanding Production data Information Analytics Analytics Knowledge (what, how, why) Support decision making As a routine
- 9. INTERNAL We have done improvements in several parts of our process based on analytics. • The recipient matching 70% => 90% • Shorten the response time of low-quality name-address-data (early warning system). • Adding missing data to the reference DB • Ongoing improvements of the matching algorithm. • Better API for printing houses. • Matching results used to optimize material flows. • Improving co-operation with mailers and printing houses. The most important change has been within virtual team members. • We do analysis • because we want to understand. • because we want to be better in our work. • because increased knowledge decrease uncertainty and ease the decision making. • to do less mistakes. 13.9.2023 9 Some practical results Marcus Gustafsson
- 10. INTERNAL In traditional information system development reporting and analytics are quite often seen as a non-functional requirements ~ not so important. Quite often data analytics are not a core part of the information system development. Quite often the information system developers are not part of the data analytics. And still we are talking about using information systems to support the data driven management and decision making. This division of development can’t continue if we are aiming for a data driven management. 13.9.2023 10 Concept: IS and analytics development in co-evolution Marcus Gustafsson
- 11. INTERNAL 13.9.2023 11 Developing IS and data analysis in co-evolution as a routine Analysis Information system Event data New knowledge => improvements For core functionality and analytics Doing better analytics As a routine Marcus Gustafsson
- 12. INTERNAL Questions and comments 13.9.2023 12 Thank you Marcus Gustafsson