← Return to program

Unpacking the geospatial engineering toolbox – an overview of data science techniques for spatial data

Friday 10:00 AM–10:30 AM in Hall C

Part of the All Things Data! specialist track

This talk is an introductory dive into the world of spatial data in data science. We will discuss how we use python to massage and process objects and buildings' shapes and locations, in order to solve problems such as determining accessibility to buildings and roads, bulk processing of spatial geometries at scale, or asset maintenance optimization. Whether you are a seasoned GIS professional or new to spatial data, this talk offers insights, potential speed-ups, and new project ideas. Come along, even if you think spatial data is not your thing – it is an interesting corner of data science you will want to know about!

See this talk and many more by getting your ticket to PyCon AU now!

I want a ticket!

Data Science frequently deals far more with algorithms and data processing than statistical modelling, especially if you are in a field with limited data or special formats. One of these is spatial data, commonly seen if you work in infrastructure and other engineering fields. Shapes and locations of the objects that shape our lives such as buildings, properties, lots, roads, fences, vehicles, etc. are the source of both our insights – where should we place new schools to maximize impact, or what properties may present challenging accessibility problems – and the source of our pain – because geometries, just like any other data, are messy and irregular; they take up a lot of memory with thousands of coordinates; they require conceptually different approaches compared to the more ubiquitous tabular, text, or image data; and there are many algorithms out there that you need to be familiar with to solve your issues.

This talk is a guide through the world of Python libraries and techniques for handling spatial data using real examples we solved in our daily work, such as determination of accessibility, asset maintenance optimization and bulk processing at high computational speed and with good code quality. While some of these will be familiar to people in the GIS field, we, as data scientists, are more interested in incorporating geospatial data at scale, leveraging automation where we can in our data processing and ML pipelines. If you work with spatial data in any capacity, this talk may give you some ideas for algorithms, speed-ups or even new project ideas you can try, and if you don’t see yourself needing to work with spatial data, come join to learn more about this part of data science!

Christine Seeliger

Christine is the Technical Lead of WSP Digitals Data Science and Analytics Team. A scientist by training, she has worked across a wide variety of different fields from computer science and software engineering to biomedical research, fintech and now civil engineering. She is continuously looking for new and interesting challenges to expand her knowledge and skills.

Long Dang

Long is a junior data scientist at WSP, who likes all things math, deep learning, and computer science. I am most at home crunching through some algorithm/coding problems with too much scratch paper and a pen. When I am not at work, you can find me watching anime, playing gacha games, going on random walks if the weather is nice, meeting friends and new people, or trying to get a neural network's loss to go down.