Develop a Data Mart Implementation Plan
With a clear and complete functional specification, a knowledgeable resource can configure a Data Mart very quickly. This makes it worth the time to create a good implementation plan at the start.
Inventory the Fields and Groupings to be Supported
The first step in planning a Data Mart is to take as complete an inventory as possible of all the fields and groupings you plan to support in your information delivery project. The more complete the inventory, the better you can create a logical and economical design of models, model extensions and groupings. This does not mean that use of the Mart is an all or nothing or big bang step, only that within the scope of the data to be supported in the Mart, a complete inventory of it makes for a better plan.
This inventory process should include an effort to standardize where possible on a minimal set of definitions of common fields like market value and security price. Standardization makes possible a compact Mart where data elements are widely shared among consumers of the data.
Start Small and Build on Your Successes
While it is best practice to create a complete inventory of fields you will need for a specific information delivery goal, that goal may be a modest one, especially if it is your first Data Mart project. The mart is architected for future growth. So doing a complete analysis of a present known need does not require foreknowledge of all future reporting requirements.
Identify and Create any Missing Meta Data Elements
Every Data Mart field requires a field attribute, and every Data Mart grouping requires some set of field attributes, range rules and dictionaries (regular and performance). The Mart implementation plan should identify these meta data elements quite specifically, and indicate where something must be created.
Use the Data Mart Planning Template
Global Professional Services has a template that is very useful for storing all of the Data Mart and regular meta data elements involved in a Mart, and can serve as the basis of an implementation plan.
Plan a Sources Strategy
A sources strategy for Data Mart is a design that uses the minimal list of source rules and source-specific fields required to provide the entire source perspectives needed in reporting. Often, this implies a single source rule if data of every type, such as positions and trades, can correctly be reported using just a single hierarchy of sources. More than a single source rule is required only if certain individual positions, trades, or securities must be viewed in more than one way within the scope of information delivery.
Common use cases for multiple source rules or source-specific fields include:
Multiple providers of bond ratings or other security analytic data whose viewpoints must all be considered in the security evaluation process;
Multiple accounting bases, such as GAAP and statutory, that require accounting fields to be computed and reported in different ways.
When more than one type of data must be reported in more than one source hierarchy, the number of required source rules can proliferate and lead to several snapshots and heavy duplication of data. For this reason, you should make choices to limit the required list of snapshots. See “Sharing of Security Details Among Snapshots” for additional information.
A best practice is to build best data in PACE where possible, and then use the resulting Eagle PACE source as the main source for those data types built.
0 Comments