Multi-Select Filtering with Streamlit and Snowflake

Read Time:2 Minute, 0 Second

Streamlit, coupled with Snowpark, provides an excellent platform for creating interactive and user-friendly data applications. This blog highlights the significance of “Multi-Select Filtering with Streamlit and Snowflake”, demonstrating how multi-select filters can simplify data exploration by allowing users to apply multiple filters dynamically.

While the example discussed is a basic implementation, it serves as a valuable learning tool for newcomers, like myself, to grasp the power of multi-select filters and Snowflake’s robust data querying capabilities

Imagine you’re a data analyst for a global retail company. Your company operates across multiple regions and offers a variety of product categories. Your stakeholders need a dashboard to explore sales data interactively. The requirements include:

Filter Sales Data:
- By Region: Select one or multiple regions (e.g., “North America,” “Europe”).
- By Product Category: Filter specific product lines (e.g., “Electronics,” “Furniture”).
Limit the Results: Provide a way to restrict the number of records displayed to avoid overloading the interface.
Display Results Dynamically: Allow users to see filtered results interactively with a button click.

This dashboard will serve as a beginner-friendly application but provides hands-on experience with multi-select, Streamlit, and Snowpark.

Technical details:

1. Import Necessary Libraries

# Import required packages

import streamlit as st

from snowflake.snowpark.context import get_active_session

from snowflake.snowpark.functions import col

2. Define the Layout and Filters

The layout includes:

A sidebar for multi-select filters and result limits.
A main content area for displaying filtered results.

# Initialize the layout

def init_layout():

st.title("Regional Sales Analysis Dashboard")

st.markdown("""

**Explore Sales Data**

Use the filters in the sidebar to select specific regions and product categories.

A simple application for learning multi-select and Snowpark integration.

""")

3. Fetch Unique Filter Values

@st.cache_data

def get_distinct_values(session, table_name, column_name):

query = f"SELECT DISTINCT {column_name} FROM {table_name}"

result = session.sql(query).collect()

return [row[0] for row in result]

def query_snowflake(session, table_name, filters, limit=10):

df = session.table(table_name)

for column, values in filters.items():

if values:

df = df.filter(col(column).isin(values))

return df.limit(limit).collect()

Conclusion

This Streamlit and Snowpark-powered application demonstrates the significance of multi-select in data filtering. It’s a beginner-friendly project, but it lays the foundation for building more advanced dashboards