If you’re looking for a quick and easy way to leverage analytics focused on a specific topic, look no further. During his BLUEPRINT 4D session, Patrick Wheeler, product management, Oracle Database, showed how to build a cloud data mart in just 20 minutes.
What Is a Data Mart?
So, what is a data mart? A data mart is a simple form of data warehouse focused on a single subject or line of business. With a data mart, teams can access data and gain insights faster because they don’t have to spend time searching within a more complex data warehouse or manually aggregating data from different sources. It’s exactly what a departmental analyst may want—self-service business analytics that are too big for a spreadsheet but for which an enterprise data warehouse would be overkill.
So, why might you want a data mart? Simply because you have a problem—you need to understand your business in detail.
How to Build a Cloud Data Mart
Creating a cloud data mart is a simple, five-step process:
- Prepare a workspace for the data analysis.
- Gather the raw materials (load the data).
- Transform your data into the format you need.
- Analyze the data to better understand your business.
- Share your analysis with others in your organization.
In his session, Wheeler walked through a brief demonstration, which is explained here:
To prepare your workspace in Step 1, you’ll need to provision the data platform. Oracle Autonomous Database is the ideal platform for this work. In the OCI Console, click on create an ADW Database. The workload type is Data Warehouse, and you will provision it on shared Exadata infrastructure. Determine your OCPU count and whether to enable autoscaling. This will take only a few minutes to set up. Then, you will need to connect to database actions. Using the Database Users tool, you can create different roles that can access the data warehouse.
In Step 2, you’ll need to load or link to data. In the Data Load tool, the first step is to specify how to reach it in terms of a cloud storage location. Then, you can load the specific files from that location into your data warehouse. A card will be created for each table that you load.
For Step 3, you’ll need to transform your data. You can aggregate your data to better suit the needs you have—whether it is a high-level overview or a deep dive. Wheeler’s demonstration showed how you can set up data flows to better aggregate your data to suit your needs and store it in a new table.
Once your data is set up in the format that you need, you can move on to Step 4—analyzing your data to better understand your business. Oracle Database has a structure called Analytic View that can help. In Wheeler’s demonstration, he used the Data Analysis tool to create an analytic view. The tool will guide you through the process of set up. It scans the data in the database to identify candidate dimensions, hierarchies, and measures and will provide a proposed star schema design. You can edit or rename any of the proposed hierarchies or measures in the design. Once that is set up, you can go to the Analyze tab to begin your analysis. Determine your columns, rows, values, and filters to view your data in a graph or table.
The final step is Step 5—sharing your analysis with other members of your team. This will help everyone get on the same page and begin leveraging the analytics that you’ve set up. These insights will help your team better understand the data and develop plans for improving your business.
For more information about how to build a cloud data mart and to watch Wheeler’s full demonstration, check out the full BLUEPRINT 4D presentation at https://questoraclecommunity.org/learn/ recordings-presentations/your-self-service-data-warehouse-should-be-autonomous.