#data viz

[davesanalytics]

Percent of US Women give birth by age group.

(source: Our World InData)

(Plot created using gnuplot)

[transport-methodology-101]

Mercedes-Benz 300 Messwagen, 1960.

Messwagen translates to ‘measuring car’ and that’s what this one-off Mercedes-Benz is. A wagon created for the purposes of collecting data on other cars.

Built in 1960, the Measuring Car is based on a Mercedes-Benz 300 chassis. It packs a 3.0-litre dual-carburetted straight-six producing 118kW. It needed all of those kilowatts to move around equipment that was simply too heavy and cumbersome to move and assemble otherwise. Despite being a unique solution to a tricky data-gathering problem, Mercedes-Benz only built O N E single example of the Messwagen. They wouldn’t build another fully in-house station wagon until 1977.

https://www.carsguide.com.au/oversteer/weird-wagons-mercedes-benz-300-messwagen-59645

Mercedes-Benz 300d Versuchsbegleitwagen, 1960 - mb143.ru

Mercedes-Benz 300d Versuchsbegleitwagen, 1960 - mb143.ru

[throuthewindow]

This week on "I Love This Thing So Fricking Much," we spoke with Joel, who loves data and creating data visualizations that help teams achieve their goals. Find out why by tuning in on Spotify, Apple Podcasts, the link above, or wherever you listen to podcasts!

We are actively looking for more guests to tell us about the things you love! Reply to this post, tag a passionate friend, or use the podcast guest submission form linked in our bio.

[solarpunks]

Olafur Eliasson Designs a Canonical Structure with 832 Vibrant Glass Panels That Reflect Sonoma’s Weather

Drawing on the microclimate of the vineyard, the studio constructed the mosaic of translucent and transparent panels using meteorological measurements of solar radiance, wind intensity, temperature, and humidity. A winding gravel path leads to the outdoor seating area, and as the sun passes over the area, it drenches the brick construction in a full spectrum of color, a contrast to the Northern California landscape.

Ream more at thisiscolossal.com

[data-has-feelings]

Knitting for the quietly enraged

Yarn-based data visualisation – which is very much a thing, if you didn’t already know – often seems to come with an undercurrent of rage attached to it.

One of the first examples I remember was a woman in Germany, who knitted a colour-coded scarf that recorded every time train delays ruined her commute.

Not too long after came a scarf from a city councillor knitting her way through meetings in Montreal. She changed colours based on the gender of the person speaking: green for women and red for men. The scarf was, inevitably, mostly red.

These both feel allied with the slightly bigger phenomenon of the ‘temperature scarf’. Lots of people have done these, and they’re the ones you’re most likely to have seen. Artist and author Josie George knitted one in 2020, recording the weather every single day for a year.

Of course, the weather rarely behaves as it should these days. After recording a week of unseasonably hot May weather, she tweeted:

“For me, a month of deep grief and deeper resolve, side by side, as I looked unflinchingly at the world's damage. I read and spoke words of hope and change as I knit. I continued. I committed. I began again.”

Knitting is steady, continuous, and pattern-driven. It’s perfect for making a record of things as they happen.

And that’s how it goes with a lot of things we get angry about. They happen once - a train is late, a man drones on, the temperature lurches into something alarming - but it’s just one little thing. Not enough to justify rage.

And then they keep happening, and happening again. React to any one instance and you look irrational. But keep a record, quietly, politely, and calmly, and maybe you get to feel like you’re building the evidence base for a revolution.

Or not. Whether or not you get to make your point, whether or not other people listen, just the act of recording might feel important. You get to process the things that you have to put up with, without ignoring the part of yourself that’s insisting it isn’t right.

[destinationtoast]

How to address non-binary characters in fandom stats about gender

@once-a-polecat asked a thoughtful and interesting question in response to my post about Canon gender representation & shipping:

So, I know this is a VERY small number of characters (statistically speaking), but how do you deal with, or plan to deal with, the increasing numbers of characters who are non-binary in canon? (Jim Jimenez in OFMD, or Desire in Sandman for instance.) If you’re documenting how prevalent it is that fans introduce gender diversity to binary characters at some point you have to address the question of canon characters who do not have a binary gender. And I understand how much of a tricky question this is because it’s less easily traced across fandoms and some characters may not have a binary gender in ways that do not track with Western human society (they may have culturally significant genders or be non-human beings etc…), but small numbers and diverse types of genders gets difficult for data visualization purposes.

I wanted to give a long-ish answer (or really, to discuss a number of possible answers, all of which sometimes apply but aren't complete on their own), and I also wanted to open it up to others for ideas, so rather than reply in the notes, I'm replying here.

One answer is to say that I'm limited by the data I have available about canon (e.g., in that past 2018 work I was pointing to about Gender representation in canon vs. fanworks, I was working with someone else's movie data set, which didn't contain any data about canon nonbinary characters). That raises a question for future analyses -- does anyone know of a reliable list of canon nonbinary characters that is kept up-to-date? How good is Wikipedia's List of fictional non-binary characters, e.g.? Also interested in lists of canon trans characters.

(A quick aside, because I'm about to talk about limitations: I'm incredibly grateful for the AO3 tagging system, and everyone who makes it work! In part because it allows me to do far more complex stats about all this stuff than any other fandom platform -- or other media platforms, period. And because it gives us all some pretty outstanding sorting and filtering superpowers. Hallelujah! Okay, now onto some limitations. :) )

Another possible answer is to say that my stats are about how people tag things on AO3. So for the most part, I just follow however people are tagging things on AO3 -- and if AO3 users start tagging more ships as "Other" as they often do when a canon nonbinary character like Jim Jimenez or Desire is involved, some of my tag-based stats will pick up that kind of thing (even though I'm currently investigating F/F and M/M specifically, about which more below, I do often include more shipping categories than that). However, this is also only somewhat satisfying, since tagging practices differ between fandoms and change over time (e.g., I *think* more people used to tag some ships involving nonbinary characters, like LaFontaine/Perry from Carmilla, with gendered tags like "F/F" rather than with "Other" -- though maybe that particular example was just because L/P was often a background ship). And I know the use of "Other" can also be touchy; it can be frustrating to lump together human-only ships like Oluwande Boodhari/Jim Jimenez with ones like Eddie Brock/Venom Symbiote, for one thing. (And AO3 could even change some of these tagging options in the future, which might make such data even less useful.)

Another answer is to say, "Nonbinary characters are usually just noise, statistically, because there are so few of them -- I can ignore them most of the time." (Which you nodded to in your question, though you didn't suggest being so dismissive about it. :) ) There are some times when I make those kind of omissions or oversimplifications, though I avoid it when possible. For the purposes of my upcoming "F/F vs. M/M" analysis, this answer and the previous one both come into play -- my initial goal is to look at how explicitly queer fic differs between fanworks tagged "F/F" vs. "M/M." Which will unfortunately miss a whole bunch of queer fanworks featuring nonbinary characters. But I am going to simplify things by initially focusing on those two largest and most explicitly queer ship tags.

A final answer is to say that I've tried to make up for limitations in my other work by explicitly seeking out and analyzing gender diversity in some of my stats, like my analyses of Trans, nonbinary, and gender-diverse characters on AO3. [Edit: and then I addressed the possible visualization complexities by looking at each of the common tags -- like "Nonbinary Character" -- separately.] At the same time, because I don't know which characters are trans or nonbinary in canon, those stats have been vague about which gender diversity is coming from canon vs. from fandom -- see my above question about good lists of canon gender diversity. (Also, this doesn't capture types of gender diversity that aren't reflected in the tags.)

Finally, a shoutout to @centrumlumina who does a fantastic job hand-labeling the gender and race of all the characters for the annual top AO3 ships analysis. I am in complete awe. Canon race and gender are both things I wish were easier to analyze/find data about -- but at least for the characters in the top 100 ships each year, Lulu has provided a great data source!

Thanks for the question -- curious to hear if/how others think about this topic.

[datascienceunicorn]

Signal Processing Tools

HT @dataelixir

[atoms4ca]

Data Source: FERC Form 1, with ancillary data from EIA Form 860. Cost account definitions here. Visualization my own.

I have chosen to present the data on the basis of power (kW) instead of energy (kWh) because basically all of nuclear power's operations & maintenance costs are fixed (i.e. don't go down if the plant operates less).

Nuclear power plants are indeed cheaper to operate per unit of energy produced because their fuel is much cheaper (excluded from this visualization) and they operate at 90%+ capacity factors. In plain English, that means they are very close to "always on," thereby churning out more electricity per year than an identically sized fossil plant. In contrast, fossil-fired plants are first to be powered down when demand falls because (1) more expensive fuel and (2) combustion causes greater wear-and-tear, requiring more maintenance. Nuclear power plants don't have that problem.