Tools for Every Team

1. Tools for Every Team

Welcome to our fourth and final chapter in this course. In the previous video, we covered how to work with unstructured data. Now we will explore the tools in Snowflake that match different data teams' needs.

2. Three different use cases

At Snowy Peak, the data team has grown and everyone needs something slightly different from Snowflake. The data engineer wants to run a Python churn model on millions of subscription rows. The data analyst wants to explore last night's session logs using SQL and Python in the same place, without switching tools. And the product manager wants access to a live dashboard that can be run and reviewed every morning. These users have very different use cases. Let’s see what Snowflake can offer each of these users.

3. Snowflake Notebooks

Snowflake Notebooks lives within Snowsight - so it requires no additional installation or setup. A notebook gives you three different cell types (similar to Jupyter notebooks): Markdown for annotation, SQL for writing queries and Python for everything else.

4. Snowflake Notebooks

One thing that makes them particularly useful: you can define a variable in a Python cell and reference it inside a SQL cell using Jinja syntax. This is two curly braces around the variable name. This means you can write the query, annotate your findings, and add a visualization - all in one place. Notebooks need an active virtual warehouse to run executable cells. Start the session, and the compute kicks in automatically. This is the tool for the analyst investigating those session patterns where they can blend SQL and Python in a single workflow.

5. Snowpark

Snowflake offers a developer library called Snowpark - it’s available for Python, Scala and Java where the code executes inside of Snowflake. The idea is that code goes to the data, not the other way around. This means there's no moving data out to a local environment or external infrastructure to manage. That's what makes Snowpark right for production work: nightly ETL pipelines, batch ML transformations, custom functions at scale. If Snowy Peak's data engineer needs to score churn risk for every active subscriber at 2am, this is the tool for them!

6. Streamlit in Snowflake

You may have come across Streamlit in Python before, Streamlit in Snowflake is a way of building interactive apps that can run inside Snowflake. It lets you build out your app in Python, deploy within Snowflake and provide access to anyone on the team. This means all the data stays in the platform, you’re building something you can share with other team members.

7. Which tool, when?

How do you decide which tools to use? Well when you’re exploring data and mixing SQL, Python and Notes in one document - Notebooks is your best choice. If you’re developing production code such as data pipelines and models that run reliably at scale then use Snowpark. Finally, when you need to build interactive apps that can be shared with your team then utilise Streamlit.

8. Let's practice!

Put your knowledge of the different tools into practice!

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.