Back during my University years I became interested in the development of IoT
devices and worked on a couple of projects in that vein. In the same period, I
purchased both a Raspberry Pi and an Arduino Uno boards. While I did use the
first on and off as a Linux box and tried to set up a media server with it,
the latter sat almost untouched for many years.
In the lull I’ve experienced after finishing my PhD, I suddenly got interested
in the topic again, and decided to buy a few sensors to build something more
complex than a series of LEDs that turn on every other second.
In this series of posts, I will report the various challenges, issues (because,
of course, there were plenty of both) and results of my experience with
prototyping boards.
Notes on plots and dataframes Before we begin, a couple of notes on plots dataframes: this post has been prepared by converting a jupyter notebook into a markdown file using nbconvert. As it turns out, the default output format used by nbconvert to display dataframes isn’t rendered correctly by the Hugo backend I’m using for the blog. I was not able to find a “clean” solution for it, so for this post (and until I’ll find the actual fix) all tables are represented in raw html, rather than in a more decent format. I also need to figure out a better way of saving the plots, which allow to specify a facecolor, so that figures have a white background both in light and dark mode. For the time being, it’s possible to switch to light mode to fix that particular problem.
...
To begin with, let’s consider the material I am working with, to have an idea of what can be done with it and what might be needed to build something clever.
First off, my board is a Raspberry Pi 2 Model B (2015), which features neither a Wifi nor a Bluetooth connection. This did cause me quite a few headaches, since connecting wireless turned out to be more complex than I thought.
...