It may be an understatement that quality decisions are based entirely on the data quality going into the Snowflake data warehouse. Critical decisions are made across the enterprise every day as organizations rely more and more on outcomes powered by Snowflake analytics. As data driven enterprises move from transactional supported decisions into contextual based information integrity matters even more.
Contextual driven decisions require different types of data that provide additional related facts around the core data. The Snowflake data exchange provides a very powerful capability to share contextual information to support better decision making. Sometimes benchmark and/or third party data is acquired via the Snowflake data marketplace to further compliment the data analytics tools.
As data crosses environments cascading through decision making processes the need to ensure data integrity increases dramatically across the board. The responsibility for Snowflake data warehouse quality falls squarely on the shoulders of the administrator. Standard ETL boundaries limit the ability to work with data at the same level as Snowpark or Java UDF’s. The native Snowflake connector triggered by the Java UDF program gives the administrator data super powers to solve this challenge.
How to use native Snowflake connector for Put It Forward via the UDF program is extremely easy. Put It Forward has designed a set of Snowflake UDF programs that can be directly invoked within the console. With a simple command improve the decision quality being powered by the Snowflake data analytics tools at scale.
select pif_max('dataset1', 'NextBestCustomer1');
Figure 1: How to invoke Snowflake UDF program sample
What is triggering the UDF program or Snowpark is to bring the automated data science capabilities of Put It Forward via the native Snowflake connector into the data warehouse at run time. The data processing and AI powered intelligence layer provides the deep insights before the data analytics tools start their work.
Figure 2: Snowflake Data Warehouse Governance With Put It Forward
As increasing volumes of data are matched, merged and normalized for the Snowflake data warehouse management approaches must change. Simplifying the potential complexity increase. Solutions that impact decisions being enabled by Snowflake analytics increase in importance while simultaneously the volume of working data increases. This requires even higher levels of data integrity.
End to end data life cycle management is a common pattern for using Put It Forward with the Snowflake data warehouse. For the purpose of this discussion focus is on two scenarios that require AI analytics and high levels of data quality.
Core to the success of marketing is establishing meaningful rapport with the customer through engagement. This is also very true for strategic approaches such as account based marketing (ABM). Attached to this but outside of the scope of this article are Snowflake analytics scenarios like cross channel attribution.
The nature of customer data is that it’s both transactional and contextual. Making it a moving target in terms of how it’s used in data analytics. Said another way, as customer data is provisioned and managed through Snowflake data sharing, the type of use determines the outcome.
Data governance processes around Snowflake data warehouse need a different approach because of the massive volumes and variety of data types. Put It Forward leverages continuous AI and machine learning to identify the right data for marketing scenarios including the following:
As Put It Forward churns through data it’s being integrated into the Snowflake data warehouse. When this data is surfaced within Snowflake analytics it becomes more important to target the data in the right direction. Enter the Snowpark power capabilities. With a single Snowflake UDF program you can quickly work with the data while staying within the local environment.
While this scenario is focused on finance it equally applies to many other situations within the enterprise.
Finance very often has to work with third party data from multiple vendors and sources that are outside of its influence. Yet in many situations it is obligated if not regulated to deliver decision surety. Said another way, finance can’t control the quality of its inputs but is bound to the quality of its outputs. A wicked problem on a good day and a disaster on a bad day.
Financial data comes primarily in three flavors on its way to the Snowflake data warehouse.
First is transactional data that originates from internal systems. Make a withdrawal or execute a trade type of data. Second is third party data that covers reference, master and market data types. Third is data that deals with risk, portfolio management or settlement data. This could be first, second or third party sourced data.
This is a lot of highly volatile data by nature as it changes often and its context shifts simultaneously.
When operations risk management success is the criteria, working with this data via etl or data preparation processes is very difficult. Common scenarios that are used with this approach are:
Put It Forward works with the wide variety of data types, volumes and rate of change going through it via algorithms. For the Snowflake administrator or data analysts needing to work with this data it becomes a very simple process. To use it is as simple as invoking some Scala or Java code in the Snowpark environment. .
This allows the administrator to know that the data has the right attributes and integrity before propagating it into the environment. Propagation can be done with a Snowflake UDF program or integrating via Snowpark with Java or Scala. This allows you to have confidence in Snowflake analytics solutions while opening the door for monetization via the Snowflake data marketplace.
Every day some of the best organizations in the world choose Snowflake and Put It Forward to uplevel their data game. Organizations such as Gallo, BB&T, Dell and OpenTable which demonstrate leadership in customer experience for both the consumer and professional across the globe.
Everything you need to know about Snowflake integration from using no code solutions to the risks of ETL approaches.
A how to guide in approaching Snowflake data governance that lays out the needed foundation for success.
No code native end to end Snowflake integration, data governance and AI based data quality control in a single data platform.