Browser (1) Report server with compiled and deployed reports (2) OLAP engine with a cube (4) Data warehouse (where aggregated OLTP data goes to become the source for the cube) (3) and finally OFFLINE some report builder (5) Usually (2) (5) belong to a single vendor like Tableau or SSRS server/Power BI Report Builder or Pentaho The are usually compartible with most OLAP engines like MSFT Analytical services and many many other snowflake implementations For 3 and 4 we use the Mictosoft solution sql server/Analytical services And finally we chose a free reporting tool to prresent those BI reports - Power BI Report Builder
Our (any!) reports show data. Clearly we use test (automatically special program generated) data. We cannot use real data which i am sure you will agree with since you wold not want us to use your(!) data for our demos. So each client needs to fill that data warehouse table with that client's data. After that everything is ready as i explained. The specification for that data is very clear (since your stuff will have our test data as an example) and very similar to denormalized state or statistical reports. The specification is very similar to the CA SAL/SAP reports. We will also provide pseudo-code samples for SAL/SAP business logic and flow. So they (IT people) will need to write (or modify existing since you are already in the personal auto insurance business)
OLAP Engine
can you substitute some expensive OLAP engine like Oracle Essbase or maybe the opposite Open Source OLAP engine (Mondrian, ClickHouse) - they all suppoort MDX out of the box so you just regenerate the cube to them and then re-point Reports' connection to a new endpoin
what if you decide to change the report tool (both design and runtime/report server)? remember that the IP is in the reports business logic which is in queries/datasets So if you are seduced by some tool like Tableau cool features like dragging dimensions you just (for each report) hit new Report on that new expensive tool, copy dataset(s) from Power BI Report we have, and then create the same Tablix (matrix or table) the old report has in the new report and map the same positioned cells to that dataset's fields.
Report Tool
Despite all the sleekness of that dragging those dimensions i have to note that generally UWs and actuaries know which views are the most important for them like producers on rows and quarters and volumes on rows etc etc In Power BI it is acomplished by doing several Publixes within the SAME report off the same dataset and switching between them using a dropdown switch Yes each Publix (table, view) takes two hours at most but it is one time deal!
Greg Lipman: support@greglipman.com
Sebastian Bucur: sales@greglipman.com
Book a demo today!
