Analytics and Reporting
Analytics and Reporting
It was not long ago that analytics meant a choice between simple data reports or a complex statistical package, but the last several years have produced a confusing abundance of options.
When evaluating these options it is useful to remember that data structures that provide the best transactional efficiency are inefficient for big data analytics operations and that analytical operations can easily lock out business imperative transactions. In practice, this means that analytics are either restricted in scope or they operate on restructured copies of the data. Either approach imposes compromises on the business and its users.
Other core factors are the degree that the end user can (and will) define their own reports versus the ongoing cost of a report writing team; the computations that can be applied outside of basic database operations; installation complexities of client software or distributed cubes; charting capabilities; and the balance between fancy pre-formatted multi-panel reports (ala dashboards) versus simpler reports with dynamic progressive discovery.
The breadth of analytical options define different points of balance between these compromises - the one best for you depends deeply on your business needs. It also depends deeply on your implementation environment because a weaker tool may outperform a more advanced tool if the weaker tool has better interoperability with other tools such as the ones that define the data model and data loading.
We suggest to our clients that they treat analytic projects as exercises in compromise and that the compromises be creatively matched to the ever migrating business needs.
Live Reporting
There is an increasing need for some reporting to be on live data that is changing while the user watches. Letting the user run their own queries against live operation data is dangerous. If live reporting is a business necessity, please talk to Metabyte about our live web update technology which allows delivery of well defined measures into dashboards, operational screens and spreadsheets.
It was not long ago that analytics meant a choice between simple data reports or a complex statistical package, but the last several years have produced a confusing abundance of options.
When evaluating these options it is useful to remember that data structures that provide the best transactional efficiency are inefficient for big data analytics operations and that analytical operations can easily lock out business imperative transactions. In practice, this means that analytics are either restricted in scope or they operate on restructured copies of the data. Either approach imposes compromises on the business and its users.
Other core factors are the degree that the end user can (and will) define their own reports versus the ongoing cost of a report writing team; the computations that can be applied outside of basic database operations; installation complexities of client software or distributed cubes; charting capabilities; and the balance between fancy pre-formatted multi-panel reports (ala dashboards) versus simpler reports with dynamic progressive discovery.
The breadth of analytical options define different points of balance between these compromises - the one best for you depends deeply on your business needs. It also depends deeply on your implementation environment because a weaker tool may outperform a more advanced tool if the weaker tool has better interoperability with other tools such as the ones that define the data model and data loading.
We suggest to our clients that they treat analytic projects as exercises in compromise and that the compromises be creatively matched to the ever migrating business needs.
Live Reporting
There is an increasing need for some reporting to be on live data that is changing while the user watches. Letting the user run their own queries against live operation data is dangerous. If live reporting is a business necessity, please talk to Metabyte about our live web update technology which allows delivery of well defined measures into dashboards, operational screens and spreadsheets.
