Solving Complex Spatial Analysis Problems with QGIS

Ujaval Gandhi

Spatial Thoughts LLP

ujaval@spatialthoughts.com

QGIS Open Day

August 2023

​

https://bit.ly/spatial-analysis-challenge

CC BY-4.0

Watch Video Recording

Spatial Thoughts

• A learning platform for modern geospatial technologies.
• Free and open learning content for QGIS, Python, GDAL and Google Earth Engine.
• Learning materials used by 1 million+ users every year.
• All material is licensed under CC-BY
• Online academy for cohort-based instructor-led classes.
• Personalized mentoring and support
• Trained participants from over 100 countries.

#SpatialAnalysisChallenge

• A friendly challenge to inspire the geospatial community to spark creativity and enhance their skills.
• Challenges are inspired from real-world issues.
• Participants are encouraged to use any software, programming language, database or systems they are familiar with.

​

The Past Challenges

• Extract Buildings with Holes
• Name a Line Layer with Closest Points
• Connect buildings to Nearest Point on the Road
• Count Points in Overlapping Polygons
• Variable Buffers
• Conditional Spatial Join

​

See the archive

In Today’s Talk

• Extract Buildings with Holes
• Name a Line Layer with Closest Points
• Connect buildings to Nearest Point on the Road ✅
• Count Points in Overlapping Polygons
• Variable Buffers ✅
• Conditional Spatial Join ✅

​

• A Bonus Challenge ❓

Challenge 1:

Connect Buildings to Nearest Point on the Road

Problem Statement

​

Connect each building's road-facing edge to the nearest point on the street segment.

Datasets:

• Building Polygons
• Road Network Lines

Sample Data

Solution 1: QGIS Processing Algorithms

• Convert each building polygon to lines and extract the centroid of each edge.
• Connect each centroid with the roads layer
• Extract the line with the shortest distance

Full Solution [video][model]

Solution 2: QGIS Expressions

• Salvatore Fiandaca solved the problem in just 2 steps using QGIS Expressions
• Used the functions with shortest_line() and overlay_nearest()
• The resulting expressions can be implemented using Geometry Generators or using Geometry by Expression algorithm

Full solution [github]

Other Solutions

See the original Twitter Thread and LinkedIn Post with all solutions

PostGIS

RStats / Python / QGIS Expressions

Challenge 2:

Creating Variable Buffers

Problem Statement

​

Buffer each point in Layer A so that the buffer contains at exactly 5 points from Layer B

Datasets:

• 2 Point Layers

Sample Data

Solution: QGIS Expressions

• Abhimanyu S came up with a clever QGIS expression to compute the distance to the 5th neighbor for each point
• Expression uses overlay_nearest() function to find the the 6th neighbor.
• Use the distance() function to compute the buffer distance.
• Expression is used in the ‘Buffer’ algorithm.

​

Solution: QGIS Expressions (Improved)

• Michaël Galien suggested a PostGIS solution that is an improved version of the previous solution.
• Instead of buffering by the distance of 5th neighbor and potentially missing the 5th point, we can compute distance between 5th and 6th neighbor and buffer by the average value.
• We can adapt the previous QGIS expression to use this method.

Full Solution [video]

Other Solutions

See the original Twitter Thread and LinkedIn Post with all solutions

QGIS

RStats / Python

Google Earth Engine / FME

Challenge 3:

Conditional Spatial Join

Problem Statement

​

Add the name of the place within the polygon having the highest population

Datasets:

• Urban Area Polygons
• Populated Places Points

Sample Data

Solution 1: Geometry Generators

• Totò Fiandaca used QGIS Expressions and Geometry Generators to dynamically find and label the point with the highest population.
• Used nested overlay_contains() to find all intersecting points and then array_max() to pick the one with the highest population.

Full solution [twitter]

Solution 2: PyQGIS Processing Algorithms

• A optimized join algorithm that makes use of a spatial index.
• Find all intersecting point features within each polygon.
• Sort the features by the chosen attribute (population) and pick the feature having the highest value
• Copy the required attribute (name) from the selected feature

Full Solution [video][script]

Solution 3: PostGIS

• PostgreSQL provides a DISTINCT ON expression that can be used with SELECT to remove duplicate and keep the first row of the result.
• This makes it very easy to write a SQL query to find all intersecting points, sort them by population and keep the first one.
• We can run this query via PostgreSQL execute SQL algorithm in QGIS Processing Toolbox.

Other Solutions

See the original Twitter Thread and LinkedIn Post with all solutions

PostGIS

QGIS Modeler / GeoPandas

Bonus Challenge:

Constrained Nearest Neighbor Analysis

Problem Statement

​

Connect each school to the nearest college in the same administrative region

Datasets:

• School points
• College points
• Administrative boundary Polygons

Sample Data

Connect and Participate in the challenge!

​

Twitter @spatialthoughts

LinkedIn linkedin.com/company/spatialthoughts/

​

Summary

• ✅ QGIS offers a lot of functionality out-of-the-box!
• ✅ Explore the Processing Toolbox and Modeler.
• ✅ Expressions are very powerful. Learn to use aggregate expressions!
• ✅ Use PyQGIS to extend the core toolset.
• ✅ Connect QGIS with PostGIS to scale your analysis

​

​

​

Contact:

ujaval@spatialthoughts.com

​

Free QGIS learning materials

https://spatialthoughts.com/

​