Pragmatic methodology for BI and Data warehouse Development Pragmatic applies a iterative approach to deploy data warehouse and BI solutions,. The entire process is divided in to multiple stages. Project Planning In this phase we prioritize the requirements of features of the solution for deployment based on the following factors High Impact low cost and risk explicitly stated priorities of the business dependencies between features The objective is to get as much information assets into the hands of business users, where it can provide the most value as quickly as possible. 2. Data Warehouse Design This is a critical phase where we need to assess looking at various data sources to see a feasibility of Data Warehouse. In some cases if the Intelligence needs to be derived from a single source, Pentaho Schema Workbench can be used to directly server the Reports from the staging database itself. Also reports can be created directly from flat files If data volumes are very large and data has to be aggregated from multiple source, putting it in a Data Warehouse is considered ideal. This not only helps in better performance but can also help bring data to a single source for further analytics and archiving. Data Warehouse design is very critical as it needs to capture all facts and dimensions necessary to bring out the expected Analytics. Here are some points to remember while designing a good Data Warehouse. The primary goal of a Data Warehouse should be speed and usability Data Models can only be developed once requirements are clear The Data warehouse will constantly evolve as per new requirements 3. ETL and Data Quality The Extract Transform and Load (ETL) is a set of processes, procedures and code that can be reused in a customer environment. In order to create a Data Warehouse, we need to design ETL jobs in a planned and efficient manner. ETL deals with: In this phase we prioritize the requirements of features of the solution for deployment based on the following factors consistent method and content from different data sources into the data warehouse environment Error handling in case of exceptions Data Quality enables cleaner and validated information making its way to the Data Warehouse. If data has many quality issues it will further lead to incorrect Analytics. 4. Designing of Dashboards and Reports Once the data is captured in the right models, visualization is the next step. Dashboard and Report designer tools can be used to create various kind of visualizations. Different kind of visualizations from same data sources can be created for decision making. Mobile BI can also be enabled based on the needs.