I recently stumbled upon a very interesting post in medium from Bill from Dev Mastery. His first newsletter in my inbox contained someone from the past (about 3 years ago): the beloved “Uncle Bob” . Uncle Bob is always worth our time, he is an absolute delight to watch and I might have the opportunity to see him live this September! For now… enjoy 🙂
So, I googled “visualization madness” and it seems that I am not alone in this! Good to know.
Take a look at this blog that collects, well.. visualizations that make no sense: http://viz.wtf
It has become one of my favorites and I have been enjoying myself with its contents more than I should! But it feels good to know the struggle is real.
Take care y’all!
Good things first: I have recently landed a Machine Learning Python Developer job. It is what I’ve been wanting for a long time, to be involved with ML full time and not only on my free time and rarely at work. It is a great job, with great colleagues. It’s been about 5 months now that I have been designing experiments, instrumenting multiple runs with different parameters and algorithms, aggregating the results etc. (Yes yes, you can feel the but coming… ) Ok, great so far. My parameters are correct… My process is right… It could be optimized regarding execution time, but this won’t be a problem for now. I will get back to this in the next run…
Everything seems perfect so far.
Where’s my problem? Well.. Describing what I’ve done and what’s important in plain english. And those
damn frustrating charts. Anything that is not if then else, e.g. giving my team the sum up of what I’ve done in a “oh-I-see” kind of way, is not easy for me (yet). It is a work in progress. I ‘ve gotten a bit better (presentations vs documents : 1 – 0). But since my first impulse when I want to learn about something is to google thing-I-need-to-know tutorial, when it comes to Data Science and visualization, it seems that there’s not such a guide. At least not one I have stumbled upon. I know that visualization is a form of art highly dependent on what your goal and your means are, so I think this is expected, especially in the case of ML. So, this is the beginning of my attempt to gather any relevant material I come across and group them under easy to find tags, and hopefully get better along the way.
Everything good must start with xkcd – sums it up perfectly: https://xkcd.com/1138/
Good to know what to avoid and how not to cheat – emphasis on not – : https://flowingdata.com/2017/02/09/how-to-spot-visualization-lies/
Inspiring examples: https://dzenanhamzic.com/visual-gallery/
And my favorite sin 🙂 (It took a long time to really get the hang of it but d3.js is great! Can be difficult to debug but it gets more manageable with time.): https://d3js.org
I haven’t tried any of these but they may come handy to someone: https://www.dashingd3js.com/d3-training
Side note: I find that a good way to learn about a topic is to start answering questions on Stack Overflow. (This is probably not true for data visualization but, I’ll let you know when I try) Start from the easiest ones, work your way to more advanced and always be kind.
I recently stumbled upon a foreign (to me) term “Ligatures” and had to know what could this mean in the programming area, so this came up: https://medium.com/larsenwork-andreas-larsen/ligatures-coding-fonts-5375ab47ef8e#.hawse9tre
And this article about techniques to use your brain in full mode when coding: