Milan's Data Science Insights #011

Hi Everyone,

After a few extra issues, I am returning to the regular style. The format it still shaping, but one thing is sure - you can expect loads of geospatial goodies this year as well! Let's jump in.

1. Courses

I am super delighted to let you know that my second official LinkedIn Learning course is out - Geospatial Raster Data Analytics in Python with the following learning goals:

• Accurately describe the characteristics of raster data and differentiate it from other geospatial data types.
• Acquire, read, and visualize raster data from various sources, including population and elevation data, using appropriate software and tools.
• Perform basic data exploration techniques, including histogram analysis and log scaling, and interpret the results to enhance data visualization.
• Manipulate raster data by successfully performing operations on data sets such as clipping, resampling, and reprojecting.
• Combine multiple single-band raster files into a multiband raster file, conduct analytical exercises on the resulting data, and interpret the outcomes.

This course builds on my previous course, so make sure to check that as well - it is called Geospatial Data Analytics Essential Training.

2. Top Posts

In case you are not a regular on LinkedIn, I will share the highlights of my week here as well.

• A huge collection of time series analytics tools by Maryam Miradi, PhD
• A coding video sneak peak to my bran new raster analytics course
• Once you mastered raster data, I recommend to check this remote sensing analytics overview by Ulrik Martensson
• I also collected 25 of my Python tutorials on various geospatial data science topics.
• While I am certainly voting for Python, this collection of QGIS plugins by Mashford Mahute is a must for everyone using that software
• Speaking of Python - this colletction, aptly titled 'Life is short, Iuse Python' by Alex Wang certainly has a point and compiles a great list of must-have tools for many different data science subdomains.

3. Article Digest

As you might have seen, I called myself out to the #365Papers challenge, meaning I will follow a great advice I got as a PhD student many years ago: make sure to read at least one paper per day. Now I flip this around a bit and share one paper per day here on LinkedIn, so you can easily follow this learning journey.

• 𝐀𝐧𝐲𝐒𝐚𝐭: 𝐀𝐧 𝐄𝐚𝐫𝐭𝐡 𝐎𝐛𝐬𝐞𝐫𝐯𝐚𝐭𝐢𝐨𝐧 𝐌𝐨𝐝𝐞𝐥 𝐟𝐨𝐫 𝐀𝐧𝐲 𝐑𝐞𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧𝐬, 𝐒𝐜𝐚𝐥𝐞𝐬, 𝐚𝐧𝐝 𝐌𝐨𝐝𝐚𝐥𝐢𝐭𝐢𝐞𝐬 by G Astruc, N Gonthier, C Mallet and L Landrieu: “ultimodal model based on joint embedding predictive architecture (JEPA) and resolution-adaptive spatial encoders, allowing us to train a single model on highly heterogeneous data in a self-supervised manner.”

• A large-scale assessment of how close global cities are to the magic 15-min: 𝐀 𝐮𝐧𝐢𝐯𝐞𝐫𝐬𝐚𝐥 𝐟𝐫𝐚𝐦𝐞𝐰𝐨𝐫𝐤 𝐟𝐨𝐫 𝐢𝐧𝐜𝐥𝐮𝐬𝐢𝐯𝐞 15-𝐦𝐢𝐧𝐮𝐭𝐞 𝐜𝐢𝐭𝐢𝐞𝐬 by Bruno et. al via @Nature Portfolio
• Object detection in urban settings is getting really hot. I just came across this nice tutorial on Towards Data Science by M. Aboutaleb 𝐂𝐨𝐮𝐧𝐭𝐢𝐧𝐠 𝐨𝐧 𝐘𝐨𝐮𝐫 𝐂𝐮𝐬𝐭𝐨𝐦𝐢𝐳𝐞𝐝 𝐃𝐚𝐭𝐚𝐛𝐚𝐬𝐞 𝐟𝐫𝐨𝐦 𝐀 𝐭𝐨 𝐙
• A comprehensive overview on spatial indexing by Ben Feifke: "𝐆𝐞𝐨𝐬𝐩𝐚𝐭𝐢𝐚𝐥 𝐈𝐧𝐝𝐞𝐱𝐢𝐧𝐠 𝐄𝐱𝐩𝐥𝐚𝐢𝐧𝐞𝐝: 𝐀 𝐂𝐨𝐦𝐩𝐚𝐫𝐢𝐬𝐨𝐧 𝐨𝐟 𝐆𝐞𝐨𝐡𝐚𝐬𝐡, 𝐒2, 𝐚𝐧𝐝 𝐇3" on Towards Data Science
• The paper of the day is on mobility data analytics - 𝐌𝐢𝐱𝐢𝐧𝐠 𝐈𝐧𝐝𝐢𝐯𝐢𝐝𝐮𝐚𝐥 𝐚𝐧𝐝 𝐂𝐨𝐥𝐥𝐞𝐜𝐭𝐢𝐯𝐞 𝐁𝐞𝐡𝐚𝐯𝐢𝐨𝐮𝐫𝐬 𝐭𝐨 𝐏𝐫𝐞𝐝𝐢𝐜𝐭 𝐎𝐮𝐭-𝐨𝐟-𝐑𝐨𝐮𝐭𝐢𝐧𝐞 𝐌𝐨𝐛𝐢𝐥𝐢𝐭𝐲 by Sebastiano Bontorin et al.

• Psychology and geospatial data: 𝐑𝐨𝐛𝐮𝐬𝐭 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞-𝐛𝐚𝐬𝐞𝐝 𝐦𝐞𝐧𝐭𝐚𝐥 𝐡𝐞𝐚𝐥𝐭𝐡 𝐚𝐬𝐬𝐞𝐬𝐬𝐦𝐞𝐧𝐭𝐬 𝐢𝐧 𝐭𝐢𝐦𝐞 𝐚𝐧𝐝 𝐬𝐩𝐚𝐜𝐞 𝐭𝐡𝐫𝐨𝐮𝐠𝐡 𝐬𝐨𝐜𝐢𝐚𝐥 𝐦𝐞𝐝𝐢𝐚
• A colorful article by Robert Simmon titled 𝐅𝐫𝐨𝐦 𝐒𝐩𝐚𝐜𝐞 𝐭𝐨 𝐒𝐭𝐨𝐫𝐲 𝐢𝐧 𝐃𝐚𝐭𝐚 𝐉𝐨𝐮𝐫𝐧𝐚𝐥𝐢𝐬𝐦 shows how data journalists use remote sensing data, from wildfires, deforestation, hurricanes, to battlefields.

All the best,

Milan