Options
Data Mining Software System Development Process - Grad Student Query
mathew_pierce
Member Posts: 1 
Contributor I
Hello everyone,
I am a CS gradutate student in my first year of grad school. I'm doing a research paper on project management of a data mining tool software project, specifically for the medical radiological image arena. My project researches the history of data mining, special problems faced by data mining tool projects, and finally recommends a project management model for a data mining software system project.
I've got about 16 years of experience in the embedded system and system support arena so I've had lots of experience on projects in general but not in data mining. Since the competition in data mining tools is very fierce (stand alone tools as well as built-ins such as those provided by Microsoft and Oracle) I believe that the development process will have to be fairly agile, but not XP-ish. Since the personnel developing the tool will be some pretty deep thinkers I believe that the development process will have to be rather rigorous and somewhat rigid, in order to maintain order in the chaos. I also believe that when we come out of this current recession we'll be in a different age of job/work creation and retention; I believe there will be fewer "jobs" per se, and more "work" (ie - contract workers) therefore the project process would have to be able to accomodate people entering and leaving the business.
Having the above in mind, I'm leaning toward Rational Unified Process (RUP) because it is both somewhat ridgid in its process, allows different processes to be used to get the job done (ie, waterfall, agile, etc.) as well as being an iterative and/or incremental process. In addition, I believe that having a standard process will make it easier to have employees enter and exit the project at different stages as well as make future enhancements easier.
Does anyone have an opinion on the above?
Thanks,
Mathew Pierce
I am a CS gradutate student in my first year of grad school. I'm doing a research paper on project management of a data mining tool software project, specifically for the medical radiological image arena. My project researches the history of data mining, special problems faced by data mining tool projects, and finally recommends a project management model for a data mining software system project.
I've got about 16 years of experience in the embedded system and system support arena so I've had lots of experience on projects in general but not in data mining. Since the competition in data mining tools is very fierce (stand alone tools as well as built-ins such as those provided by Microsoft and Oracle) I believe that the development process will have to be fairly agile, but not XP-ish. Since the personnel developing the tool will be some pretty deep thinkers I believe that the development process will have to be rather rigorous and somewhat rigid, in order to maintain order in the chaos. I also believe that when we come out of this current recession we'll be in a different age of job/work creation and retention; I believe there will be fewer "jobs" per se, and more "work" (ie - contract workers) therefore the project process would have to be able to accomodate people entering and leaving the business.
Having the above in mind, I'm leaning toward Rational Unified Process (RUP) because it is both somewhat ridgid in its process, allows different processes to be used to get the job done (ie, waterfall, agile, etc.) as well as being an iterative and/or incremental process. In addition, I believe that having a standard process will make it easier to have employees enter and exit the project at different stages as well as make future enhancements easier.
Does anyone have an opinion on the above?
Thanks,
Mathew Pierce
0
