Topic: AnalyzerBeans when and how?
AnalyzerBeans when and how?
Hi DC people,
I was just reading Kaspers blog-entry about the new AnalyzerBeans "sandbox" project (http://kasper.eobjects.dk/2009/06/introducing-analyzerbeans.html) and it made me wonder how this would manifest itself into the DataCleaner application and GUI. If you merge the now separate profiling, validation and comparison functionality - how will this affect the user interface? Will it just be something "underneath the surface" or will you reflect the seemingly fundamental changes in a new UI? I don't know a lot about Java code but I have some ideas as to the user interface so I was wondering how you are going to handle these changes - and do you have any projections as to when AnalyzerBeans will be ready for use?
I was just reading Kaspers blog-entry about the new AnalyzerBeans "sandbox" project (http://kasper.eobjects.dk/2009/06/introducing-analyzerbeans.html) and it made me wonder how this would manifest itself into the DataCleaner application and GUI. If you merge the now separate profiling, validation and comparison functionality - how will this affect the user interface? Will it just be something "underneath the surface" or will you reflect the seemingly fundamental changes in a new UI? I don't know a lot about Java code but I have some ideas as to the user interface so I was wondering how you are going to handle these changes - and do you have any projections as to when AnalyzerBeans will be ready for use?
Hi franklin,
so far the AnalyzerBeans project is just an elaborate "experiment", but I would be lying if I said that I thought it would stay that way for long...
In general I see the project as an opportunity to (re)think data analysis from a clean slate. I would love to hear about ideas for user interface in the regard. Do you imagine the three "parts" (profiler, validator, comparator) of DataCleaner blending together? Do you imagine a whole new kind of interface? Would it be better to move for a web-based interface? Or would it be even better to go for integration with ETL tools? Or would that narrow the scope of the tool for "just" ETL-work...?
These are some of the questions that I think are interesting in regards to the UI of AnalyzerBeans.
so far the AnalyzerBeans project is just an elaborate "experiment", but I would be lying if I said that I thought it would stay that way for long...
In general I see the project as an opportunity to (re)think data analysis from a clean slate. I would love to hear about ideas for user interface in the regard. Do you imagine the three "parts" (profiler, validator, comparator) of DataCleaner blending together? Do you imagine a whole new kind of interface? Would it be better to move for a web-based interface? Or would it be even better to go for integration with ETL tools? Or would that narrow the scope of the tool for "just" ETL-work...?
These are some of the questions that I think are interesting in regards to the UI of AnalyzerBeans.
If the changes are as fundamental as I can recon from the description of AnalyzerBeans, I think a rethought UI is needed.
First of all I think it will be cruzial to merge the three parts as you call it. This is obviously a benefit IMHO.
Secondly I think a more flow-oriented view is required if it will be possible to chain one analyzer's output to another analyzer's input. I'm thinking something similar to most ETL user interfaces but maybe a bit more simplistic.
Thirdly I think you should provide a way to implement new analyzers within the UI itself. Of course this will require some coding skills, but I don't see why users shouldn't be allowed to do the Java(?) coding directly in the application. Or maybe some language with a bit more narrow scope?
First of all I think it will be cruzial to merge the three parts as you call it. This is obviously a benefit IMHO.
Secondly I think a more flow-oriented view is required if it will be possible to chain one analyzer's output to another analyzer's input. I'm thinking something similar to most ETL user interfaces but maybe a bit more simplistic.
Thirdly I think you should provide a way to implement new analyzers within the UI itself. Of course this will require some coding skills, but I don't see why users shouldn't be allowed to do the Java(?) coding directly in the application. Or maybe some language with a bit more narrow scope?
Seeing that we are now using analyzerbeans.jar in the DC lib folder, I will congratulate you on archieving this project :)
Log in by clicking the login link at the top of the screen
Go back to forum.
