@villares#matplotlib is cool, but I don't think it is good to make interactive stuff with it, #pygame is cool but interaction with #py5 that inherits its setup/draw/events structure from #Processing, to me, feels even easier. Maybe it is because I'm so used to it... Anyway, check the code for this shapely exploration above and make your own conclusions!
@piko
It's hard to get to the point of knowing what you're doing in #matplotlib, because they obfuscated a very well-structured API with lots of high level methods that keep breaking the API layers all the time.
It's easy to get almost the result you want. And usually it's a complete rebuild from scratch to get from this point to a good result.
If you manage.to get to the point where everything is done with low level methods, tweaking gets quite easy and efficient.
I've been playing with #OpenStreetMaps via #osmnx, which is awesome , but I struggle with simple stuff like adding a bunch of places as markers. Everything looks a bit like the owl drawing meme, either showing something too easy and useless, or something too advanced and also useless or beyond my comprehension. Maybe some other Python tools?
(I know about Marcelo's fabulous PrettyMaps but it is not exactly a viz tool)
Did you know that you can create stunning tables in Python? 🔢 📊
I used #MakeoverMonday data on Highest Paid Athletes to explore Plottable - an awesome #python library to customise tables. Really like the inbuilt graph options and how easy it is to display images. 🤩
I know #matplotlib can do cartography (maps), and #osmnx uses it to plot stuff, but I can't find documentation or a decent tutorial to plot a base map and a list of latlong places as different sized circles. I'm struggling with #cartopy, in theory a wrapper to make mapping easier :((
Is it that hard or I'm just too dumb and/or I'm making everything wrong? #Python
After weeks of making #maps for the 30DayMapChallenge, it's time for some #dataviz again.
Looked at what words C-3PO uses most often together (bigram analysis) and visualised them in a network graph (thicker line=more co-occurences). "Oh dear" 😃
Change in population of London 1990-2020 per 1skm. Growth almost everywhere but largest in East London. Unsurprising considering new developments around Stratford (ie. Olympics) and other areas in East since then.
Chciałbym także podziękować autorom PyPy za ich wsparcie, zarówno w kwestii poprawiania błędów w PyPy, jak również udzielaniu pomocy innym projektom, by poprawić ich zgodność z PyPy. Praca z wami jest przyjemnością!
Na koniec, poznałem ważny argument za pracą nad wsparciem PyPy w projektach: nawet jeśli dana paczka nie działa szybciej na PyPy, to może być zależnością w większym projekcie, w którym PyPy ogółem przynosi lepszą wydajność.
Forest area per country. 🌳🌲Was surprised to see such a low number in Iceland, apparently it was 40% before the Vikings arrived 😯 More here https://www.iceland.org/geography/forest/
Some in particular use #matplotlib, which I used in the distant past. If memory serves, it could handle large datasets, but "large" today might be an entirely different order of magnitude. Are you talking billions of datapoints, that sort of thing?
What resource limit(s) are you running into? Memory for Jupyter, for your JS/browser Sankey tool? Something else? If Jupyter, have you increased the default memory limit?
Annoyed by this #matplotlib default behavior. The easy (easiest?) workaround is to change the year values to text strings, but I'd like the default tick locator to be a bit smarter about this.
Of course, you can fairly easily calculate your own relative size in those, but still. In most cases relative font sizes would be more foolproof for charts and graphics than absolute sizes in points.
For research projects where I use #NumPy and #MatPlotLib, I actually like using #Jupyter! It is just easier to run code and view my plots.
With #PyQt, you can even have widgets like sliders and other #Qt stuff.
It just speeds up my prototyping and makes me more productive. Naturally, only my plotting code and math exist in the .ipynb, and the rest is just imported from normal .py files. Thus, it allows for quick conversions once the prototyping is done.
It is a vertical polar bar chart rather than a horizontal polar bar chart, but getting closer faster that I thought!
Btw, has anyone created a grouped horizonal/radial polar bar chart in #matplotlib ? I have some ideas on how to create it, but right now they all seem too complicated.
Oh, and the chart shows the #democracy status for Spain in 2023 if you are curious
I had a hunch that I could take on tasks that I used to do with Python + matplotlib + Jupyter with Swift Charts + SwiftUI + Playgrounds instead. But I had no idea it’d be this nice and easy. Plus I find the default result better looking and I have much more control over everything around the chart, like labels, titles etc.
This will change everything for me when working with charts.
I've had more opportunities recently to create plots with #SwiftUI / Swift Charts / #Swift playgrounds and the process absolutely holds up in replacing #python / #matplotlib / #jupyter.
The dataset was larger this time, ~100k records, each containing a handful of data points, exported from a Postgres DB as JSON.
Massaging the data in a typesafe way is an absolute blessing – it's so easy to plot the wrong thing in python as you drill through three dicts.