Getting Started With Snowpark for Python and Streamlit

What is Snowpark?

Snowpark at its core provides an API that developers can use to construct DataFrames that are executed lazily on Snowflake's platform. It enables data engineers, data scientists, and developers coding in languages other than SQL such as Python to take advantage of Snowflake's powerful platform without having to first move data out of Snowflake. This enables data application developers to run complex transformations within Snowflake while taking advantage of the built-in unlimited scalability, performance, governance and security features. Learn more about Snowpark.

What is Streamlit?

Streamlit is a pure-Python open-source application framework that enables developers to quickly and easily write data applications. Learn more about Streamlit.

What You'll Build

A web-based data application that uses Snowpark for Python (public preview) and Streamlit

What You'll Learn

How to create a Session object for connecting to Snowflake
How to create Snowpark DataFrames and load data from Snowflake
How to display tabular data and interactive charts using Streamlit
How to run the web-based data application using Streamlit
Tips and tricks for enhancing the web-based data application in Streamlit

Prerequisites

Familiarity with Python
Python 3.8
Snowpark for Python library
A Snowflake account with ACCOUNTADMIN role
- For this guide, we'll use the Environment Data Atlas dataset provided (for free) by Knoema. In the Data Marketplace, click on Get Data and follow the instructions to gain access to KNOEMA_ENVIRONMENT_DATA_ATLAS.
- In particular, we will analyze data in schema ENVIRONMENT from tables EDGARED2019, WBWDI2019Jan, and UNENVDB2018.
Streamlit Python library

Install conda to manage a separate environment by running: pip install conda. NOTE: The other option is to use Miniconda
Create environment using Snowflake Anaconda channel by running: conda create --name snowpark -c https://repo.anaconda.com/pkgs/snowflake python=3.8
Activate conda environment by running: conda activate snowpark
Install Snowpark for Python including compatible Pandas version by running: pip install "snowflake-snowpark-python[pandas]"
Install Streamlit by running: pip install streamlit

NOTE: You many need to use pip3 instead of pip

Let's start by creating a Python script and adding the import statements to include the required libraries.

my_snowpark_streamlit_app.py

# Import required libraries
from snowflake.snowpark.session import Session
from snowflake.snowpark.functions import avg, sum, col,lit
import streamlit as st
import pandas as pd

In this step, you'll create a Session object to connect to your Snowflake. Here's a quick way of doing that, but note that hard coding credentials directly in code is not recommended in production environments. In production environments a better approach would be to load credentials from AWS Secrets Manager or Azure Key Vault, for example.

# Create Session object
def create_session_object():
   connection_parameters = {
      "account": "<account_identifier>",
      "user": "<username>",
      "password": "<password>",
      "role": "<role_name>",
      "warehouse": "<warehouse_name>",
      "database": "<database_name>",
      "schema": "<schema_name>"
   }
   session = Session.builder.configs(connection_parameters).create()
   print(session.sql('select current_warehouse(), current_database(), current_schema()').collect())
   return session

In the above code snippet, replace variables enclosed in "<>" with your values.

In this step, you'll create three Snowpark DataFrames to load data from tables EDGARED2019, WBWDI2019Jan, and UNENVDB2018 from schema ENVIRONMENT.

# CO2 Emissions by Country
snow_df_co2 = session.table("ENVIRONMENT.EDGARED2019").filter(col('Indicator Name') == 'Fossil CO2 Emissions').filter(col('Type Name') == 'All Type')
snow_df_co2 = snow_df_co2.group_by('Location Name').agg(sum('$16').alias("Total CO2 Emissions")).filter(col('Location Name') != 'World').sort('Location Name')

# Forest Occupied Land Area by Country
snow_df_land = session.table("ENVIRONMENT.\"WBWDI2019Jan\"").filter(col('Series Name') == 'Forest area (% of land area)')
snow_df_land = snow_df_land.group_by('Country Name').agg(sum('$61').alias("Total Share of Forest Land")).sort('Country Name')

# Total Municipal Waste by Country
snow_df_waste = session.table("ENVIRONMENT.UNENVDB2018").filter(col('Variable Name') == 'Municipal waste collected')
snow_df_waste = snow_df_waste.group_by('Location Name').agg(sum('$12').alias("Total Municipal Waste")).sort('Location Name')

In the above code snippet, we're leveraging several Snowpark DataFrame functions to load and transform data. For example, filter(), group_by(), agg(), sum(), alias() and sort().

More importantly, note that at this point nothing is executed on the server because of lazy evaluation–which reduces the amount of data exchanged between Snowflake and the client/application. Also note that when working with Streamlit we need Pandas DataFrames and Snowpark API for Python exposes a method to convert Snowpark DataFrames to Pandas.

# Convert Snowpark DataFrames to Pandas DataFrames for Streamlit
pd_df_co2 = snow_df_co2.to_pandas()
pd_df_land = snow_df_land.to_pandas()
pd_df_waste = snow_df_waste.to_pandas()

As mentioned above, the Snowpark DataFrames are lazily evaluated, which means the SQL statement is not sent to the server for execution until an action is performed on it. An action, for example to_pandas() in our case, causes the DataFrame to be evaluated and sends the corresponding generated SQL statement to the server for execution.

At this point, you're technically done with most of the code and all you need to do to render the data in a web application in your browser is to use Streamlit's dataframe() API. For example, st.dataframe(pd_df_co2).

But let's add a few more web components to make our data application a bit more presentable and interactive.

In this step, you'll add...

A header and sub-header and also use containers and columns to organize our dataframes using Streamlit's columns() and container()
Tabular display of the data using Streamlit's dataframe()
Interactive bar chart using Streamlit's slider() and bar_chart()

# Add header and a subheader
st.header("Knoema: Environment Data Atlas")
st.subheader("Powered by Snowpark for Python and Snowflake Data Marketplace | Made with Streamlit")

# Use columns to display the three dataframes side-by-side along with their headers
col1, col2, col3 = st.columns(3)
with st.container():
   with col1:
   st.subheader('CO2 Emissions by Country')
   st.dataframe(pd_df_co2)
with col2:
   st.subheader('Forest Occupied Land Area by Country')
   st.dataframe(pd_df_land)
with col3:
   st.subheader('Total Municipal Waste by Country')
   st.dataframe(pd_df_waste)

# Display an interactive bar chart to visualize CO2 Emissions by Top N Countries
with st.container():
   st.subheader('CO2 Emissions by Top N Countries')
   with st.expander(""):
      emissions_threshold = st.number_input(label='Emissions Threshold',min_value=5000, value=20000, step=5000)
      pd_df_co2_top_n = snow_df_co2.filter(col('Total CO2 Emissions') > emissions_threshold).toPandas()
      st.bar_chart(data=pd_df_co2_top_n.set_index('Location Name'), width=850, height=500, use_container_width=True)

In the above code snippet, a bar chart is constructed using Streamlit's bar_chart() which takes a dataframe as one of the parameters. In our case, that is a subset of the CO2 Emissions by Country dataframe filtered by column Total CO2 Emissions via Snowpark DataFrame's filter() and user-defined CO2 emissions threshold set via Streamlit's user input component slider().

The fun part! Assuming your Python script (as shown below) is free of syntax and connection errors, you're ready to run the application.

You can run the by executing streamlit run my_snowpark_streamlit_app.py at the command line. (Replace my_snowpark_streamlit_app.py with the name of your Python scrupt.)

In the application:

You can click on columns to sort the data
You can increase/decrease the emissions threshold value using the slider to change the data visualization

App

my_snowpark_streamlit_app.py – Source Code on GitHub

# Snowpark
from snowflake.snowpark.session import Session
from snowflake.snowpark.functions import avg, sum, col,lit
import streamlit as st
import pandas as pd

st.set_page_config(
     page_title="Environment Data Atlas",
     page_icon="🧊",
     layout="wide",
     initial_sidebar_state="expanded",
     menu_items={
         'Get Help': 'https://developers.snowflake.com',
         'About': "This is an *extremely* cool app powered by Snowpark for Python, Streamlit, and Snowflake Data Marketplace"
     }
)

# Create Session object
def create_session_object():
    connection_parameters = {
      "account": "<account_identifier>",
      "user": "<username>",
      "password": "<password>",
      "role": "<role_name>",
      "warehouse": "<warehouse_name>",
      "database": "<database_name>",
      "schema": "<schema_name>"
    }
    session = Session.builder.configs(connection_parameters).create()
    print(session.sql('select current_warehouse(), current_database(), current_schema()').collect())
    return session
  
# Create Snowpark DataFrames that loads data from Knoema: Environmental Data Atlas
def load_data(session):
    # CO2 Emissions by Country
    snow_df_co2 = session.table("ENVIRONMENT.EDGARED2019").filter(col('Indicator Name') == 'Fossil CO2 Emissions').filter(col('Type Name') == 'All Type')
    snow_df_co2 = snow_df_co2.group_by('Location Name').agg(sum('$16').alias("Total CO2 Emissions")).filter(col('Location Name') != 'World').sort('Location Name')
    
    # Forest Occupied Land Area by Country
    snow_df_land = session.table("ENVIRONMENT.\"WBWDI2019Jan\"").filter(col('Series Name') == 'Forest area (% of land area)')
    snow_df_land = snow_df_land.group_by('Country Name').agg(sum('$61').alias("Total Share of Forest Land")).sort('Country Name')
    
    # Total Municipal Waste by Country
    snow_df_waste = session.table("ENVIRONMENT.UNENVDB2018").filter(col('Variable Name') == 'Municipal waste collected')
    snow_df_waste = snow_df_waste.group_by('Location Name').agg(sum('$12').alias("Total Municipal Waste")).sort('Location Name')
    
    # Convert Snowpark DataFrames to Pandas DataFrames for Streamlit
    pd_df_co2  = snow_df_co2.to_pandas()
    pd_df_land = snow_df_land.to_pandas() 
    pd_df_waste = snow_df_waste.to_pandas()
    
    # Add header and a subheader
    st.header("Knoema: Environment Data Atlas")
    st.subheader("Powered by Snowpark for Python and Snowflake Data Marketplace | Made with Streamlit")
    
    # Use columns to display the three dataframes side-by-side along with their headers
    col1, col2, col3 = st.columns(3)
    with st.container():
        with col1:
            st.subheader('CO2 Emissions by Country')
            st.dataframe(pd_df_co2)
        with col2:
            st.subheader('Forest Occupied Land Area by Country')
            st.dataframe(pd_df_land)
        with col3:
            st.subheader('Total Municipal Waste by Country')
            st.dataframe(pd_df_waste)
    
    # Display an interactive chart to visualize CO2 Emissions by Top N Countries
    with st.container():
        st.subheader('CO2 Emissions by Top N Countries')
        with st.expander(""):
            emissions_threshold = st.slider(label='Emissions Threshold',min_value=5000, value=20000, step=5000)
            pd_df_co2_top_n = snow_df_co2.filter(col('Total CO2 Emissions') > emissions_threshold).to_pandas()
            st.bar_chart(data=pd_df_co2_top_n.set_index('Location Name'), width=850, height=500, use_container_width=True)

if __name__ == "__main__":
    session = create_session_object()
    load_data(session)

Here are a couple of tips and tricks to note:

You can change the theme (light or dark) by clicking on the hamburger menu on the top right and then clicking on the Settings menu
Making any changes to the source script and saving it will automatically prompt you to Rerun the application in the browser without having to stop and restart the application at the command line. (This can also be configured to always rerun the app without a prompt.)
You can create multi-page apps in Stremlit.
You can use Streamlit's st.session_state to save objects like snowflake.snowpark.Session so it's only created once during a session. For example:

if "snowpark_session" not in st.session_state:
  session = Session.builder.configs(json.load(open("connection.json"))).create()
  st.session_state['snowpark_session'] = session
else:
  session = st.session_state['snowpark_session']

Congratulations! You've successfully completed the Getting Started with Snowpark for Python and Streamlit quickstart guide.

What You Learned