Our customer organizations have a goldmine - their data - that they want to leverage for existing data consumers and, as time goes on, make it available to even more business users.

A traditional data warehouse is no longer the only way to build data stores and large data solutions. Organizations need a suite of data solutions that can help to address specific use cases.

That's why Snowflake is a valuable piece of the data agility puzzle. As referenced in a previous article, Big Compass has successfully deployed Snowflake for many customers.

But, Snowflake may not be suitable for every data-oriented use case. Wondering if your use case is one that Snowflake can help solve? Here are three real-world examples of implementations that Big Compass has helped clients with that can get you started thinking about Snowflake's potential for your business.

Use Case 1: Retail Transaction Analysis

In a retail environment, transaction data comes in large quantities. But quantity isn't the only challenge for data analysis. Data must be kept fresh and current. Even with well-designed processes, your ETL and refresh may be time and resource constrained.

Even if you solve your data volume by increasing the power and size of your servers, you may have to repeat the increase in the future. Backups of the data are also a concern that may keep you worried, even with the routine and frequent database backups. Add to the data issues a growing list of current and future consumers of the data and the problems with your retail transaction warehouse expand. 

These were what one of our clients experienced. The problems hampered their ability to consume the data and made it difficult to provision information for multiple customers who needed to:

• Do retail sales analysis
• Understand seasonal impacts
• Analyze rewards programs
• Compute and report on rebate programs.

The solution we brought to them included Snowflake. The platform has several capabilities that can help address these problems, including:

Abstraction: Snowflake's processing abstraction with Warehouses allows scaling of compute power to meet the business's needs without changing the infrastructure.

Role-based access: Data in Snowflake can be accessible by role, and also allows users to mask PII data or limit available fields with Secured Views. This means that teams can see the data they need to do their work while maintaining adherence to governance standards.

Backups: Snowflake's Time Travel feature keeps 90-day backups, that are stored on a regular basis, so you can quickly revert to an older version of the data set (or even “Undrop” a table)  if anything goes wrong.

Use Case 2: Making Healthcare Analytics

Trend research can help healthcare organizations improve patient outcomes by identifying conditions, behaviors, and environmental factors. To do this research, organizations need massive amounts of data relating to public health.

Adding complication, most organizations don't rely solely on their own data. They work with other organizations and use their data, which comes in many formats. Those formats might include flat files, CSV, JSON, and even XML. It can't be uniformly processed in its current format and reformatting the data would be a huge undertaking.

Since public health information has great value, other organizations may also benefit from their views and research using the same data. Those different audiences may not be tech-savvy and could require a straightforward interface so that they, too, can analyze the data for healthcare advancements and research.

Snowflake has several features that can address these challenges. Those features include:

Datalake reads: Snowflake can read from an Amazon S3 datalake. Staging the disparate data will help organize the information and allow you to use the External Tables feature to view the data in a structured or semi-structured manner.

Data views: Using the Variant columns feature, Snowflake allows you to create semi-formated tables and load all JSON and XML data into your database, object by object. This enables you to create views that are formatted and user-friendly. 

Stored Procedures: Snowflake's extremely powerful stored procedures help you transform all of the various data from the different sources into a single format.

JDBC adapter: Using the JDBC adaptor, Snowflake allows you to connect directly to external tools like Tableau so that citizen analysts can still access and leverage the data.

Use Case 3: Fueling Machine Learning

It would be great to have a crystal ball to predict changes in the stock market accurately. While we all know a crystal ball isn't realistic, you may be able to build an intelligent solution that will help you reduce the risk of choices and improve your possibility of being right.  

To build this type of solution, you would plan to use historical stock data and trade data, news, and legislative data. You would want everything that impacts your machine learning application's analysis to help you make more accurate predictions. Unfortunately, some of this data will have to be manually acquired and curated, resulting in a less than fully automated solution.

To improve performance, the model will need to be constantly re-trained as the dataset grows and be highly responsive so that the system users have time to act on the information it produces.

Of course, as you add more data into your dataset and performance, more people will be interested in having access to and using the outputs of the solution.

Snowflake offers several features that can help address this hypothetical applications needs:

Manually uploaded data: SnowSQL allows the uploading of curated data directly into tables.

User-defined functions: The Snowpark feature lets you build user-defined functions (UDF) developed in Scala (Python and Java support coming soon) that can run natively and combine with stored procedures to shift much of the processing off your servers.

Multi-cluster warehouse: Assigning a large, multi-cluster warehouse to your team in Snowflake lets you run multiple high-load queries simultaneously with quick responses.

Monetization: The Snowflake data marketplace lets you monetize the massive and valuable dataset you've collected.

Conclusion

Not every data solution needs every platform. But Snowflake is a powerful and flexible solution that can augment your warehouses and provide scalable and accessible access to cloud-based analysis across your organization. It's not a transactional database and would be inefficient and costly for that purpose. But it is an ideal tool for working with large data loads, stratification of information and allowing tailored access to information.

Big Compass has helped companies implement these types of use cases with Snowflake. We'd love to meet with you about your data challenges and explore the right data agility solutions for your business. We even offer a half-day workshop to help you better understand the next steps you should take to achieve your goals with data, the cloud, APIs, and integrations. Contact us to discuss the next step in your journey.