Classifying the sortware development tasks : How to find good categories for software development tasks for deeper analysing
Punnek, Elvis (2024)
2024
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202501131268
https://urn.fi/URN:NBN:fi:amk-202501131268
Tiivistelmä
This Master’s thesis topic originated from the need to improve daily work done as scrum master
and team lead. And the idea came from real need to improve performance of the work.
Main objective of the work was to find ways how to classify software development tasks so their
work estimates could be improved. Outcome of improvement would come from the understanding
of work estimation mistakes done on specific categories of the tasks.
Key factor was to understand what kind of categories would need to be generated from current
tasks. This way there would be predefined categories that would have known need to more specific
valuation of the task estimate. This was done by manually going through selected set of existing
tasks and analysing what kind of categories they would generate.
Goal was to understand current material and then do research how similar problems have handled
in different cases. This would provide good information how to go forward and generate categories
in the future. Manual analysis work of the material was successful, and it gave lot of information
about the quality of current task descriptions. It also gave better understanding of the structure of
the current task descriptions.
When researching the external exiting material, it clarified the fact that this kind of classifying should
not be done manually, but machine learning should be used instead. Also, during research other
possible improvements were discovered for task handling.
Information gathered during the thesis work will give good baseline information how to go forward
with the classification. The work also gave good information about what other possibilities machine
learning could bring to task handling in future.
and team lead. And the idea came from real need to improve performance of the work.
Main objective of the work was to find ways how to classify software development tasks so their
work estimates could be improved. Outcome of improvement would come from the understanding
of work estimation mistakes done on specific categories of the tasks.
Key factor was to understand what kind of categories would need to be generated from current
tasks. This way there would be predefined categories that would have known need to more specific
valuation of the task estimate. This was done by manually going through selected set of existing
tasks and analysing what kind of categories they would generate.
Goal was to understand current material and then do research how similar problems have handled
in different cases. This would provide good information how to go forward and generate categories
in the future. Manual analysis work of the material was successful, and it gave lot of information
about the quality of current task descriptions. It also gave better understanding of the structure of
the current task descriptions.
When researching the external exiting material, it clarified the fact that this kind of classifying should
not be done manually, but machine learning should be used instead. Also, during research other
possible improvements were discovered for task handling.
Information gathered during the thesis work will give good baseline information how to go forward
with the classification. The work also gave good information about what other possibilities machine
learning could bring to task handling in future.