Welcome to the Building Geospatial Multi-Lay Apps with Snowflake and Streamlit quickstart. Today you will learn how to analyse and transform geospatial data in Snowflake. You will be using Ordnance Survey open datasets available on the marketplace as well as the worldwide open overture buildings dataset provided by CARTO.
This is a progressive learning experience from viewing points on a mapk right through to building a multi layer app - which pulls together buildings with unique property reference numbers, the road network and urban extents.
You will be covering:
- Geospatial datatypes and formats
- Points, Linestrings and Polygons
- H3
- Spatial Joins and Calculations
- Visualising the results using Pydeck
- Search Optimisation
- Building a multi-layer app - which pulls together everything you have learned.
Thoughout the experience, you will demonstrate the concepts with Snowflake Notebooks and Streamlit.
Structure of the session
The Lab consists of three notebooks and two Streamlit applications.
- Completion of the (ANALYSE_LOCATION_DATA) notebook.This is an end-to-end tour on how location data works with Snowflake
- Creating a Streamlit app ROAD_NETWORKThis is an example Streamlit application which brings together the different techniques learnt in notebook 1.
What You Will Learn
- An understanding of geospatial data types and formats - and how this works in Snowflake
- An understanding of Snowflakes built in H3 capabilities
- Spatial Joins
- Geospatial Calculations
- Geospatial Filtering
- Using Search Optimisation to speed up geospatial queries
- Visualing points, linestrings and polygons on a multi layered map using Pydeck - all within a Streamlt application
What You'll Build
- A streamlit application to visualise the UK road network together with buildings - which uses Ordnance survey road network data and the Carto Overture maps buildings dataset.
Prerequisites
- A Snowflake Account. Sign up here for a free 30 day trial.
Open up a new SQL worksheet and run the following commands. To open up a new SQL worksheet, select Projects » Worksheets, then click the blue plus button and select SQL worksheet.
CREATE DATABASE IF NOT EXISTS ANALYSE_LOCATION_DATA;
CREATE WAREHOUSE IF NOT EXISTS LOCATION_ANALYTICS;
CREATE SCHEMA IF NOT EXISTS NOTEBOOK;
CREATE SCHEMA IF NOT EXISTS STREAMLIT;
CREATE SCHEMA IF NOT EXISTS DEFAULT_SCHEMA;
- From the home page Navigate to Data Products » Marketplace
Ordnance Survey Datasets
- In the search box provided, search for the following datasets:
- Ordnance Survey - Urban Extents for Cities, Towns and Vilages
- Ordnance Survey - Postcodes, Place Names and Road Names
- Ordnance Survey - Road Network Great Britain - Open Roads
- Ordnance Survey - Unique Property Reference Numbers - Great Britain: Open UPRN
- For each dataset found, press Get to get the data.
- Follow the on screen prompts to install each dataset.
Carto Datasets - Obtain the following datasets from the marketplace
This dataset provides sample building polygons all over the world.
- Carto - Overture Maps Buildings
Are you ready to start learning about location data in Snowflake?
This tutorial will take you through how you can use location data to perform spatial calculations, joins, and visualise the data using the popular Pydeck python package. We will be using the freely available datasets which you have now installed to step through examples of how geospatial data can be handled.
- Click here to download notebook
- From the home page, navigate to Projects > Notebooks
- Import the notebook as a new Snowflake Notebook
- Choose LOCATION_ANALYTICS as the database and Notebooks as the schema
- Name the notebook ANALYSE_LOCATION_DATA
- Before you start the notebook, click on packages and import the Pydeck package
- Press Start to initialise the notebook and follow the instructions within the notebook.
This app gives you an example of how you can bring all these datasets together to form a multi layered mapping visual.
- Navigate to Projects > Streamlit
- Create a new streamlit application and call the application ROAD_NETWORK inside the LOCATION_ANALYTICS database and the STREAMLITS schema
- Delete all sample code and import the pydeck package by clicking on packages on the top left hand corner of the screen.
- copy and paste the following code into the Streamlit canvas from here
Conclusion
Congratulations on completing the Building Geospatial Multi-Layer Apps with Snowflake and Streamlit Quickstart! Throughout this session, you've explored how Snowflake can be used to analyze and transform geospatial data, combining multiple datasets to generate valuable insights.
What You Learned
- Fundamentals of Geospatial Analysis in Snowflake
You started by learning how to work with location data in Snowflake, visualizing points on a map and understanding spatial relationships. - Building a Multi-Layer Application
Using Streamlit, you explored how to bring together different layers of geospatial information into an interactive application.
Next Steps
With this foundational knowledge, you can extend your analyses by:
- Integrating additional external datasets, such as demographic or economic data, to enhance decision-making.
- Experimenting with different visualization techniques to improve data storytelling.
- Automating geospatial data processing with Snowpark and external APIs.
We hope this hands-on lab has provided you with the confidence and skills to apply geospatial analytics in your own projects. Happy analyzing! 🚀