MatecDev

An opinionated selection of the best C++ libraries for numerical and scientific computing, with a ready-to-build CMake boilerplate using FetchContent.

A hands-on comparison of nanobind and pybind11 for calling C++ from Python, with working code examples, benchmarks, and practical advice on which to choose.

We build a simple total cost of ownership model to find the break-even utilization rate at which owning your own HPC hardware beats renting from the cloud - and explore what happens when you stretch depreciation beyond the usual 3 years.

A practical guide to buying used GPUs from miners, gamers, and data centers to build a budget GPU server for CUDA development, AI experiments, and scientific computing.

In this post we compare the different GPU cloud providers and try to find the best ones for those who want to learn CUDA.

In this post we go through some important considerations on how to pick a budget GPU for CUDA development.

Get started with stock market data analysis in Python using yfinance, yahooquery, or yahoo_fin.

A quick review of some programming languages, platforms, and reasons to teach your kid how to program

Get started handling multidimensional arrays in Julia with this simple guide.

Combining the incredible flexibility of Python with Fortran for high-performance number-crunching is an excellent idea, especially if you already have some legacy Fortran code hanging around.

There are numerous ways to call FFT libraries both in Numpy, Scipy or standalone packages such as PyFFTW.

Julia has several tools to help you organize your code

The floating-point performance of various array-type structures in C++ can be confusing, as there are several options to pick from, and different ways to use them.

In this blog post, we will explore the differences between stack and heap allocation, their advantages, disadvantages, and when it's appropriate to use each one.

In this post, we'll discuss how you can indeed perform data analysis tasks without using anything beyond basic high-school math.

Julia promises high-performing code and ease of use, but also has a few downsides.

Tuples in Julia are fixed-length collections of values, of potentially different data types, which can't be modified once initialized.

Nobody expects good performance from native Python, but we do expect it when calling compiled functions in libraries like Scipy...

If you want to keep you code well organized, the first step is to create a project.

A quick overview of things to do when your Julia code performance is not what you expected.

Quick reference guide on how to get started with Plots.jl

Multiple plots can be combined together into a single figure using layouts.

A good way to plot numerical data that spans a very wide range of values, is to employ a logarithmic scale in one or both axis.

Set the basic series attributes in order to choose between solid, dashed or dotted line plots, bar plots, or scatter plots

Set the legend locatin in a convenient position, including outside the plot.

Configure or get rid of the gridlines in a plot by setting the grid parameters.

A collection of example plots with source code

Make your plot informative and readable with text and proper font sizes.

Control where and how data appears in a plot by setting limits and aspect ratios.

Structs are a great way to represent data in a compact and easy-to-understand way.

Mathematics comes into play in many areas of game (and especially game-engine) development.

Acquiring data-skills (coding and statistics) and combining them with domain-specific knowledge is increasingly more important as the field matures.

In this post, we'll discuss what are the main professional roles in the data industry, and where mathematics actually fits.

In this section we will briefly review existing packages within the Julia ecosystem for rootfinding and optimization.

The Newton-Raphson method is a rapidly-converging method to approximate roots of a smooth function.

The bisection method is the simplest rootfinding algorithm.

We present two alternatives to generate random points within a polygon in python: a very simple (but slow) method, and a faster one that relies on Geopandas spatial joins.

We can split a polygon into equal-sized rectangles that form a grid, using Shapely.

You can create a Shapely polygon from a list of points, either by providing the point's coordinates or extracting them from a list of Shapely points.

We can plot Shapely polygons by resorting to Geopandas plot() function, or directly, extracting the polygon's boundaries.

In order to merge shapely polygons, we can resort to the unary_union() function in shapely.

We can extract points from a polygon using the mapping() function, or with detail, extracting the polygon's boundaries.

Learn the most popular geospatial python libraries such as shapely, geopandas and rasterio, with this hands-on, detailed tutorial.

This post will show you how to leverage cloud-optimized GeoTiffs and the AWS free tier to access satellite imagery, using Python.

Working with legacy Fortran code can be challenging as it requires deep knowledge of several programming languages and the ability to interact with the domain experts.

Google Earth Engine (GEE) is no doubt a powerful service, kindly provided for free by Google.

Multiple dispatch is a simple idea with far-reaching consequences.