Comparing Data Flows in Oracle Analytics Cloud with Spreadsheet-Based Tools

In many companies, IT departments play a crucial role in gathering and organizing data from disparate sources. Ideally, these companies would implement processes to handle large amounts of data and would have a tool to effectively prepare data for reporting and analysis. It’s difficult to have large amounts of data formatted and aggregated properly for every reporting use case. Many companies rely on spreadsheets to prepare data for reporting and analysis, yet spreadsheets present a few challenges:

• Spreadsheets are often prone to data entry and formula errors.
• They lack visibility into the transformations that have been applied.
• They lack automation.
• They introduce security concerns when emailed.
• Spreadsheets don't support advanced statistical and machine learning capabilities.

Data flows in Oracle Analytics Cloud (OAC) enable users to leverage the benefits of spreadsheet-based tools while alleviating many of these challenges. The low-code environment makes it easy for existing spreadsheet users to build preparation workflows and to derive meaning from their data stores.

This article highlights some of the advanced capabilities provided by data flows, including the ability to track data transformations through the visual UI, automate workflow runs, share workflows with other users in a secure manner, and leverage advanced machine learning models with no coding experience.

Tracking Transformations

Data flows in OAC are designed so that each “step” implements a change to the dataset. It’s clear what steps were applied, along with the order in which they were applied. There are four main “categories” of data flow steps: data ingestion, data preparation, machine learning, and database analytics. Each of these categories offers a range of functions for data preparation. Data flows allow users to view a preview of the data after each step to verity that the transformations yield the desired result. Unlike spreadsheet-based tools, the data flow editor combined with the data preview makes it easy to track changes and observe how each transformation influences the resulting dataset.

Automation and Scheduling

Users of OAC can configure schedules on their data flows to control when and how often the data flow is run. This introduces a layer of automation to ensure that output datasets used in workbooks are up-to-date and reflect the information present in sources. You can configure data flow schedules to meet your business needs: for example, you can configure schedules to run at a certain time after your source database systems update. You can also configure schedules to run at hourly, daily, weekly, monthly, or yearly increments. Running data flows using schedules prevents users from having to manually run the workflows and introduces a layer of automation often lacking from spreadsheet tools. The following animated image illustrates how users can create and configure schedules for data flows.

Data Flow Sharing

Users in OAC can share data flows with other users and application roles, which presents several benefits related to collaboration and governance. The benefits and use cases related to data flow sharing include:

• Teams can easily collaborate on single data preparation workflows, reducing duplicated efforts.
• Sharing data flows allows for quality control and error reduction. Allowing multiple users to work on and/or have access to a data flow can potentially prevent data preparation mistakes from being overlooked.
• If someone leaves the company or goes on vacation, the data flow is accessible to other users. The company doesn't have to create data flows from scratch.
• Sharing data flows makes troubleshooting easier. If a user runs into an issue, they can share with someone else on their team for assistance.

Sharing data flows with users and application roles is relatively straightforward, as the following animated image illustrates.

Tailored to Citizen Data Scientists

If you're new to machine learning or don’t have much experience coding in Python or R, data flows in OAC offer the ability to easily prepare data and apply machine learning models in the same low-code workflow. No coding experience is required to leverage many of the data flow machine learning offerings. If you're interested in learning more about the machine learning offerings, check out this link. Many spreadsheet tools don't offer the wide range of machine learning algorithms that are provided in OAC.

Source: oracle.com