User:Intervex/PrideColours

This is a page where I talk about some original research I've been doing on colour-meaning associations in pride flags. This is a project that is under active development and information may change.

My motivations:

• I've been archiving pride flags here on Wikimedia Commons and have been wanting to figure out useful categorization schemes for them.
• I also want to make some nice infographics about the common colour-meaning associations.
• I want to improve the cognitive accessibility of pride flags by establishing the most common colour-meaning associations.
• I make pride flags sometimes and would like to know what the most common meaning associations are

I'm hoping this will be useful for:

• Archivists and people providing captions/alt text to flags, both for deciphering and for consistently describing flags.
• Flag designers, for creating flags with clear affordances.
• Flag users, for deciphering new flags.

Tables of the most common colour-meaning-associations are under Results. Probably the most important thing to know before jumping down there is that I am not using RGB or HSV or similar to do my analysis of colours. I'm using a different colour space called okLCH. In okLCH, colours are decomposed into lightness (0-1), chroma (0-0.37), and hue (0-360°). The hue values in this document are okLCH hue values, and those are different from the hue values in HSV/HSL!

Background

Pride Flags

There are so many pride flags. It can be kind of overwhelming even for ableminded people. Since developing long COVID I've found out extra hard with my brain fog and memory problems.

The MOGAI community has made some real strides in improving flag accessibility in terms of offering more sensory friendly flags and providing alt-text for images more reliably.

But I feel like we haven't really talked about cognitive accessibility.

Flags that follow common patterns are easier to understand. Patterns can arise in a bunch of ways like:

• Patterns can be in colour choices (pink means feminine, blue means masculine)
• and shading (light pink means girl, hot pink means attraction to girls, dark pink means AFAB).
• Patterns can be be in layout (e.g. ace & aro flags and derivatives).
• Patterns can be in icons used (like how a lot of romantic flags take the sexual version and add a heart)

I really really appreciate it when I see a new flag and I can figure out what it means from my knowledge of other flags. Here are some examples of flags that other people have made that I could immediately figure out:

• 
  Intersex and disabled. Swaps the intersex ring out for the disabled icon. Love it.
• 
  Trans and intersex. It's the trans flag and the intersex flag! Super clear, also love it.
• 
  Ace and kinky. The ace flag with the kink symbol. Again, clear!

Whereas some flags I've encountered I found confusing or unintuitive (at least at first). For example, despite openly identifying as bisexual for two decades, it was not until a month ago that I learnt its meanings are not "attracted to female-both-male" but actually "attracted to same gender, attracted to both/between, attracted to opposite gender".

• 
  This is same-both-opposite, not female-both-male (I got this wrong for years!)
• 
  The green in the polygender flag means abundance (nonzero) and not gender neutral.

Similarly, it was not until I started this project, I mistakenly thought the polysexual flag was for attraction to women, men, and gender-neutral people. Because I read the green as gender-neutral. It is supposed to mean attraction to multiple genders but not all genders - green means nonzero but not infinite. I now understand this as associating green with abundance (nature!).

Indeed, learning the green was for abundance helped! It helps me out when flag designers explain their designs. Not just what all the colours mean but /why/. What the thought process was. It helps me not only understand a given flag but can help me understand similar flags.

I'm hoping that establishing a standard set of colour-messing associations will help improve cognitive accessibility. One thing I went into this project thinking is that we need to do a better job of differentiating greens. Green gets used in a lot of different ways and it would help if we were consistent about lime green meaning one thing vs forest green meaning another vs bluish green meaning yet another thing (etc).

What I didn't want to do was just... write down what I think the standards ought to be. Because I could be totally off-base from other people (an all too familiar experience as an autistic).

So rather that being prescriptive my goal here is to be descriptive. To catalogue what trends are already being used and make those trends clear (where they exist).

In gathering data for this project my goal has been to log everything without judging. Like a lot of people I have some opinions on flags. I have very deliberately included flags I don't like (be it on aesthetic, conceptual, etc grounds). The only limits I allowed myself were I could skip adding flags if the user posting them was openly and unambiguously a TERF in the given post, as well as I could skip adding flags where the poster was obviously and unambiguously shitposting. Wherever I was uncertain that a flag was posted in good faith I erred on the side of including it.

I was originally going to put a glossary of genders and attraction types here, but this has gotten rather long, so have instead put it in the Results section. See Results by Column for definitions.

Colour theory: an introduction

Colour is complicated. One thing making it complicated is there's a rather big difference between Colour: The Physical Phenomenon (i.e. specific wavelengths of light) and Colour: Human Perception Thereof. Our eyebrains do all sorts of wacky things with Colour: The Physical Phenomenon. For example:

• We see colours like pink that don't have specific wavelengths (our brains see certain mixes of blue and red and... make a colour with that.)
• The cones in our eyes are not uniformly sensitive: we see some wavelengths with more sensitivity than others. (See: Colour vision)
• Different cultures/languages draw the boundaries between common colours at different places. We have an easier time distinguishing colours we have more practices differentiating. We often don't really see colour as a uniform gradient so much as a series of categories. Where you draw the line between green and blue is something that tends to have strong cultural associations. (See: Blue–green distinction in language)

How we mentally model colour also is complicated. We get different results when we talk about mixing beams of light together vs. mixing pigments/paints together.

• 
  Additive colour mixing: everything together combines to make white
• 
  Subtractive colour mixing: everything together combines to make black

Colour spaces

A colour space is a way of modelling colour. It sets up a system that can be used to relate colours to each other. This can then be used to determine what colour is "opposite" a given colour, as well as to determine what a mixture of two colours would be in this system.

Chances are, if you've learnt one colour model, it was RYB. It's an old system, developed for painting. It uses red, yellow, and blue as its primary colours, hence the name RYB. Secondary colours are created through mixing paints/pigments of the primary colours. Tertiary colours can be made through secondary-primary mixes and so on.

• 
  RYB colour, a colour model developed for mixing paints

RGB

Designing colour models for computers is a bit of a different animal than creating colour models for paint mixing. Chances are, if you've learnt one colour model for computer colour, it's RGB. In RGB, every colour is composed of how much red light there is (0-255), how much green there is (0-255), and how much blue there is (0-255). Unlike RYB, RGB uses green rather than yellow as a primary colour. What is or is not a primary/secondary/etc colour is dependent on the colour system you're using.

• 
  RGB forms a space that looks like a cube
• 
  This cube can be navigated as a three-dimensional space with R/G/B being x/y/z coordinates

When you see hex codes that represent colours, they are representing RGB: the first two letters are the red quantity converted into hexadecimal (base 16), the next two are the green quantity converted into hexadecimal, and the last two are the blue quantity converted into hexadecimal.

HSV and HSL

RGB values tend to be rather unreadable. If I tell you a colour is (202, 59, 104) can you figure out what kind of colour it would look like? I certainly would struggle, and I know I'm not alone.

Readability is a reason people tend to like HSV and HSL. These are two very similar colour spaces that take the gamut (total range of possible options) of RGB colours and convert the shape of the gamut from a cube into a cylinder. A colour in HSL space is composed of three values:

• Hue: 0-360° with red at 0°.
• Saturation: 0-1, sort of means how much grey is in the colour.
• Lightness: 0-1, sort of means how light the colour is.

• 
  In HSL, the top of the cylinder is white and the most vibrant colours are at the bottom. In HSV, the most vibrant colours are at the top of the cylinder and black is at the bottom of the cylinder.

While HSL and HSV improve the readability of RGB, they still have a bunch of drawbacks. The remainder of this section is heavily influenced by these articles by Lea Verou, Geoff Graham, Keith J. Grant, and the next several examples come directly from this article by Alexei Boronine

Problem #1: HSV value & HSL lightness have no consistent meanings

These two colours have the same lightness values in HSL, even though to human perception they look like they have wildly different lightness levels:

HSL(250, 100%, 50%)	HSL(60, 100%, 50%)
HSL(250, 100%, 50%)	HSL(60, 100%, 50%)

Problem #2: HSL/HSV saturation is also a mess

Here are two colours with the same saturation values in HSL (but different lightnesses). They do not look equally saturated:

HSL(0, 90%, 40%)	HSL(0, 90%, 80%)
HSL(0, 90%, 40%)	HSL(0, 90%, 80%)

Problem #3: HSL/HSV are not perceptually uniform

This means the hues are not spread out in a way that makes sense to human colour perception. For example, a 20° difference in hue in the yellow area of HSL/HSV is a much bigger difference to our eyes than a 20° difference in the blues:

HSL(30, 100%, 50%)	HSL(50, 100%, 50%)
HSL(30, 100%, 50%)	HSL(50, 100%, 50%)
HSL(230, 100%, 50%)	HSL(250, 100%, 50%)
HSL(230, 100%, 50%)	HSL(250, 100%, 50%)

A fundamental issue with HSL/HSV and RGB is that they were all designed around what is computationally/physically easiest to do with light and computer monitor technologies in the 1980s.

OkLCH

The good news is that we have much better alternatives to HSL/HSV and RGB!

OkLCH is a relatively new colour space that has a lot of nice features and has support in CSS, browsers, etc. One of the big things going for it is a perceptual colourspace. This means it has been designed so that it has been designed around Colour: Human Perception Thereof whereas RGB/HSV/HSL/etc were designed around Colour: The Physical Phenomenon.

A colour in okLCH space is composed of three values:

• Lightness (L): 0-100% with 100% being white. How light the colour is.
• Chroma (C): 0-0.37 though this range will likely increase with improvements in computer monitor technology. This is how vivid a colour is. A chroma of 0 is grey.
• Hue (H): 0-360°, with pink at 0°. While a similar idea as HSL, the actual values are different. Colours with a given hue in HSL will have a different hue in okLCH.

• 
  The hue distribution in okLCH: 0° is pink, 180° is teal. 90° is yellow (brown if dark) and 270° is indigo.
• 
  The same hues, but with less lightness and less chroma.
• 
  Different hues have different maximum chroma. This is because it's a perceptual colourspace - remember human colour perception is not equally sensitive to every hue. (I don't actually know what colour space this image is from but okLCH has similar sorts of peaks.)

Feature #1: Lightness means something

If two colours with different hues/chromae have the same lightness value in okLCH, they actually look equally light! For example, a yellow with the same lightness as a dark blue winds up looking dark (brown)!

oklch(59.47% 0.092 245.75)	oklch(59.47% 0.092 105)
oklch(59.47% 0.092 245.75)	oklch(59.47% 0.092 105)

Feature #2: Chroma means something

Chroma, which is very roughly similar (but not equivalent) to saturation actually has consistent results across different hues and lightnesses! These reds have the same hue and chroma, but different lightness:

oklch(47.19% 0.11 19.81)	oklch(79.19% 0.11 19.81)
oklch(47.19% 0.11 19.81)	oklch(79.19% 0.11 19.81)

Feature #3: Perceptual uniformity

A 20° difference in hue is actually the same thing to our eyes when it's 20° of blues vs 20° of yellows:

oklch(59.47% 0.092 245.75)	oklch(59.47% 0.092 265.75)
oklch(59.47% 0.092 245.75)	oklch(59.47% 0.092 265.75)
oklch(59.47% 0.092 105)	oklch(59.47% 0.092 85)
oklch(59.47% 0.092 105)	oklch(59.47% 0.092 85)

One reason perceptual uniformity matters is for making gradients. HSV/HSL has this weirdness when it comes to gradients where it bunches up around certain hues. OkLCH and other perceptual colourspaces will instead have a smooth gradient:

• 
  Comparison of HSV and okLCH with regard to a full-rainbow gradient. The okLCH is perceptually uniform, whereas the HSV has clear vertical lines which emerge.

Feature #3: More intuitive colour mixing

All of the computer-based colour spaces we've been talking about operate by treating colour as a 3D shape, with using three values to navigate the x/y/z or equivalent polar coordinates of this 3D shape.

This means we can do math with the numbers! If I want to get the midpoint of green and blue, I can take the coordinates for the green and blue and average them across each axis. So if I have a green with RGB (31, 159, 0) and a blue with RGB (91, 206, 250), I can calculate the mixture as:

• red: (31 + 91) / 2 = 61
• blue: (159 + 206) / 2 = 182
• green: (0 + 250) / 2 = 125

And then I get... a more bluish green than the one I started with! Exciting! Except... it's not a great midpoint now is it? You'd expect the midpoint between green and blue to be teal. And doing this with HSV gives me a green that looks somehow lighter than both the green and blue I started with.

But because okLCH is a perceptual colourspace, when I do the same thing in okLCH, I actually get a teal!

• 
  Colour mixing by averaging coordinates in colour spaces: okLCH actually gives us a teal when we do this with green and blue!
• 
  Doing this with pink and blue, RGB gives us grey! Both HSV and okLCH give purples, with okLCH's purple matching the lightness of the input pink & blue.

Mixing pink and blue this way gets at another problem with RGB and HSV: Dead Grey Zones. Sometimes when you make gradients between two colours in RGB and HSV you wind up with greys in the middle. Here's an example of a gradient in HSL that yields grey in the middle.

Feature #4: MOAR GAMUT

Last thing I wanna mention is gamut. Gamut is the total range of colours that are available in a given colour space.

Not all colour spaces have the same size of gamut. RGB has a larger gamut than CMYK, a colourspace used for colour printing. This means there are some colours, like vibrant greens, that we can see on our computer monitors but if we tried to print the image/website/etc to paper, these greens would get desaturated.

• 

If you've noticed that newer televisions and computer monitors seem to have more vibrant colours... it's because they literally do! Computer monitor technology has improved a lot recently.

Most monitors these days support a gamut of colours close to a standard called P3. P3 has a 50% larger volume than what traditional RGB supports.

This means that if you're using RGB for your webpages (CSS/etc), visual art, etc, you **cannot access** all of the possible colours that your monitor can actually show you.

The gains in gamut size have been overwhelmingly in the vivid parts of the colour spectrum. OkLCH supports P3 colours, meaning it gives us access to these vivid colours that are not in RGB but that computer monitors can actually show us.

On top of this, okLCH has been designed in a way that makes it future-proof.

Remember how I said Chroma ranges from "0-0.37 though this range will likely increase with improvements in computer monitor technology"? That's what I mean. Right now the maximum chroma supported by typical computer monitors is around 0.37. But there is no upper limit to chroma. In ten years it might be 1.0. In twenty it could be 2. Who knows? But the flexibility is there.

Feature #5: okLCH support

OkLCH is not the only perceptual colourspace in town. But being a newer one, it's been designed to fix minor issues with older colourspaces like CIELAB and old-school LCH.

The big thing here is that okLCH has CSS support. This means when you're making websites (or the tables in this page!) or the like, you can replace hex codes with okLCH codes. The okLCH codes look like this: oklch(59.47% 0.092 85)

okLCH is also supported by libraries like color.js and chroma.js. You can find more resources here and here.

Here's a great okLCH colour picker that will also let you do conversions between hex/RGB/etc.

Methods

I assembled a data set of colour-meaning associations in pride flags. It contains 2060 colour choices from 624 pride flags (meaning, on average, each flag had 3.3 colour choices). There are 1587 unique colours among the 2060 colour choices (77%).

The data set

Each row of the spreadsheet represents a colour choice made in a flag design. If there is a known meaning for this colour choice, the meaning is recorded in the appropriate column for this row.

In a flag like the trans flag, there are three colour choices: the pink (girl), blue (boy), and white (transition/intersex/nonbinary).

In a flag like the intersex flag, there are two colour choices: the yellow (not a gender) and purple (not a gender).

When flags come in a set, the common colour choices are shared. So the greys of the demigirl and demiboy flags are shared between them rather than repeated.

• 
  Trans flag: there are three colour choices: the pink (girl), blue (boy), and white (transition/intersex/nonbinary).
• 
  Intersex flag: there are two colour choices: the yellow (not a gender) and purple (not a gender).
• 
  Demigirl flag: there's one colour choice here: pink (girl) because the white & greys are common to all demigender flags.

Data collection

I gathered flag data over the course of multiple weeks, in multiple ways. In chronological order:

• I started by working through pages that are compilations of common pride flags with their associated meanings, such as the one produced by UBC Okanagan's equity office and the one by Outright International.

• I then searched Tumblr for "pride flag meaning" and spent a bunch of time going through the newest flags posted here with meaning descriptions.

• I went through the intersex flags I already had on hand as somebody who makes/coins intersex flags, and added ones with known meanings.

• I went through LGBTQIA+ wiki's entire category of culturally-specific genders and recorded the ones that had known colour meanings. Turns out most colour choices in culturally-specific flags are to use the colours of the relevant nationality, such as in how the Muxe flag uses the colours of the Mexican flag. (Such meanings were not tagged as culturally-specific genders, but rather references to nations.)

• 
  The green, white and red of the Muxe flag come from the Mexican flag.
• 
  The Kathoey flag is the Thai flag and the trans flag joined together.

At this point I was noticing I had very few stripes/colours for things that are not queer - for example, I had a ton of stripes that symbolize non-conformity, but very few that symbolize conformity.

So I searched Tumblr, Reddit, DeviantArt, LGBTQIA+ Wiki, and Google for flags that have meanings that contain "conformity", "binary", "straight", "heterosexual", "transhet", "queerhet", "traditional", "passing", "stealth", "questioning", and added the search results (most of which were the opposite of my search terms).

Later on, when I got to tagging for types of attractions, I realized I had relatively few flags which contained aesthetic attraction, so I searched tumblr and LGBTQIA+ wiki for aesthetic attraction and added what I found.

Similarly, when tagging for other identities, I realized I had relatively few flags for kink - Tumblr and LGBTQIA+ wiki seem to generally avoid kink flags. So I consulted fetish/kink pride flag galleries: that of Gay Fetish Goth and that of Twilight Guard.

Data tagging

I then started categorizing the meanings. I created five columns in my spreadsheet for tags:

• Gender (e.g. feminine, masculine, androgynous)
• Attraction (e.g. romantic, sexual, platonic)
• Values (e.g. freedom, openness, community)
• Disability (e.g. Deaf, blind, chronic illness)
• Other (e.g. non-Western, socialist, Jewish)

A colour choice can be tagged in multiple categories - for example, a colour choice which means "romantic attraction to women" would be tagged with "romantic attraction" in the Attraction column and "feminine/female/women" in the gender column. I did all my tagging in my LibreOffice spreadsheet - as such I did not see the actual colours while tagging their meanings, just their hex codes. (With the exception of white and black, I generally can't easily tell what a colour looks like from only its hex code.)

I started my tagging focusing on the meanings which signify a gender. I started by categorizing narrowly - e.g. "questioning" was an initial category, separate from "uncertain" and separate from "unknown". I refined categories iteratively and asked a friend for a second opinion in a handful of cases where categorization wasn't clear.

I refined my gender tagging over a period of a week - I wanted to make sure the categories felt stable in my mind. I ended up with 40 tags. For each tag, I converted all the colour choices to OkLCH space, and then took the median lightness, chroma, and hue, to yield a median colour for each tag. I made an image showing the median colours for the different tags.

Wanting to compare only the hues, I then made an alternative visualization where I set the lightness and chroma to 0.66 and 0.1. I then realized that many tags had the same hue. From this I clustered tags around common hues. My initial plan was to pick a weighted mean or median from the values in each cluster. Implementing this, I realized this was not great for spreading the hues around to be clearly distinct. Then I started tagging for attraction types and my plan went out the window, because the spread of hues was rather different for the attraction tags (more pinks!).

I decided to finish the tagging process before figuring out a hue-clustering plan. I spent another week tagging the other columns, focusing next on attraction. I added 24 tags for different kinds of attraction. I also tagged my columns for other identities. The end result is 84 tags for gender/attraction identities. (I'll post the disability & values results at a later time!)

I tried several approaches to clustering/bucketing the hues. I eventually landed on dividing the 360° of possible hues into twenty-four 15-degree buckets and rounding hues into these buckets. I wanted to be able to distinguish two types of yellow - which I've called gold and lemon - given how many yellows are in the set, so I set buckets at 90 and 105 to capture the two ends of the yellow range. The other buckets are all increments of 15° from 90° and 105°.

I made one alteration to this pattern, which is to move chartreuse from 120° to 115°. I wanted chartreuse to be more clearly in between yellow and green, and when it was at 120° I was having a hard time differentiating it from 135° ("lime").

Once I had my giant table of 80 (later 84) tags, bucketed into 24 hues, I realized there could also be overlap in lightness and chroma. So, I rounded lightness in buckets of 5%, and chroma in buckets of 1%. In cases where this yielded the same result for multiple tags, the tag with the most number of entries was given priority, and the other tag(s) were re-rounded to the next nearest lightness/chroma bucket(s).

Median hue calculation

A difficulty in calculating medians for hues is that hues are modular: a 355°-pink is closer to a 20°-red than it is to a 300°-purple.

If I have a set of hues like 350°, 5°, and 340°, I want the computed median to be 350° and not 340°.

So what I did for each set of hues was I found the largest gap between two sequential hues, and rotated the list so that the largest gap was at the end of the list. I then took the median of the rotated list (modulo 360°).

As an example:

• input list: 5, 15, 300, 320, 360
• add the first element plus 360° to the end: 5, 15, 300, 320, 360, 365
• calculate the difference between each element: 10, 285, 20, 40, 5
• the largest gap is: 285
• so the list rotates to: 300, 320, 360, 365, 375
• the median is then calculated: 360
• and if it is greater than 360, we subtract 360, yielding: 0

The convention in oklch as well as many other colour spaces is that monochromatic colours (white/grey/black) have a hue of 0°. Initially when I calculated median hues the monochromatic colours were included in the calculations. I realized this was biasing the results towards pinkish red (0°). I then changed my code to exclude the monochromatic values from the median hue calculation. (They are still used when calculating median lightness and median chroma.)

Clustering Tags by Opposites

Once I had my large table of 80 tags clustered by hue-buckets, I realized that many (but not all) tags were approximately 180° from their thematic opposites. For example, intersex and perisex (not intersex) were about 180° apart. Sexual attraction and romantic attraction were also about 180° apart.

I then started grouping tags which had their opposites in alignment. The column "set" in the table below is sets of up to four tags separated by 90° that are coordinate terms. For example, binary is 180° from non-binary and exobinary, trinary/quaternary is 90° from that, and xenogender is 180° from trinary/quaternary (anthrogender/non-xenogender). So they have been clustered into one set of four.

For another example of a set, I noticed that not only sexual & romantic attraction about 180° apart, alterous and sensual attraction were each 90° from sexual/romantic and 180° from each other. So sexual, romantic, alterous, and sensual attraction form another set.

Not all tags had quite so satisfying a set. The nearest "opposite" I could find for feminine was gender-ambiguous (rather than masculine). Feminine and masculine were 120° apart.

Clustering tags into these thematic sets yielded four more tags. I realized I had no opposite to "placeholder" (105°) so I went back to my data to see if I did have something appropriate to put in there. I noticed in my values column that my tag for something being in a current/present state had a median around 285°, so I moved this tag from the values column into the gender/identity analysis. (Post-hoc analysis to fit a pattern is a totally questionable research practice that I am justifying on the grounds that this is a fun hobby project.)

Once I'd done an initial round of opposite-clustering I noticed in my ungrouped tags that kinky/fetish was 180° from Christian. I was very amused at the prospect of putting Christian down as the opposite as kinky, and had that in my table for a while. Eventually I felt bad for the kinky Christians, so I went hunting in my data to see if I could reasonably create a vanilla tag. (I was unable to find a vanilla pride flag, though I tried.) I found an orange with the meaning "purity", a yellow with the meaning "simple and generic", and a purple with the meaning "generic", and tagged them all as vanilla, yielding a median yellow that actually looks vanilla-coloured.

Similarly, my data had no opposite for "love". When I showed my results to a friend, they pointed out that the opposite would be a green, and envy/jealousy is usually associated with green. So, I searched LGBTQIA+ wiki for "envy" and "jealousy" and found 15 flags with colour choices signifying envy/jealousy. Turns out gender envy is a thing. (Note: when I added the flags with envy/jealousy meanings, I only added the envy/jealousy stripes and not the entire flags.)

This yielded a median green of 158° for envy/jealousy that is honestly borderline between 165° (opposite of love) and 150°. But it is narrowly closer to 165° and makes thematic sense as an opposite to love so I'm keeping it there.

Finally, I realized I had no tag to act as a foil for culturally-exclusive genders. Once again I went to the existing tags in my Values column to see if there was anything suitable in the right hue range, and there was! "Universality/commonality" was approximately 90° to culturally-exclusive genders. So I pulled that tag into the gender/attraction analysis to complete that set.

These three post-hoc tags are all questionable in terms of internal validity, but they do make thematic sense. The vanilla tag is literally vanilla-coloured. The association of green and envy is a well established colour-meaning association. The "current/present" and "unviersality/commonality" tags were part of my initial tagging, just in the Values column before pulling them into the gender/identity analysis. I freely admit these tags came out of me wanting to find patterns rather than it emerging naturally from the data, and hope they make enough sense to warrant their inclusion.

Results

Results by tag

tag	n	col	set	axis	cluster	refined-L	refined-C	refined-Hue	refined-colour	refined-H	refined-H-constLC	rawH-constLC	raw-colour	raw-L	raw-C	raw-H
monogamy	2	attraction	Rel	Q	rose	0.4	0	360	#484848	#353535	#c47890	#c47890	#353535	0.329979	0	0
aesthetic attraction	48	attraction	Attr2	B	amaranth	0.65	0.16	15	#df5f6f	#e36675	#c8787e	#c8787e	#e46674	0.666758	0.15681632563553	15.2203398569623
physical attraction	12	attraction	AttrG	G	red	0.5	0.14	30	#a43c2f	#a54033	#c87a6d	#c87974	#a53f3d	0.50857	0.136148672236442	24.4142547385243
fast	26	attraction	Abs1	R	vermillion	0.7	0.14	45	#e57f4f	#e67c49	#c57e5e	#c57e5c	#e57d47	0.695943	0.146306936335122	46.3779947133055
passion / intense	27	attraction	Abs2	R	vermillion	0.65	0.15	45	#d86d38	#cd612a	#c57e5e	#c47f5a	#cb6320	0.616096	0.151948464649127	48.636806122327
sensual attraction	24	attraction	Attr1	A	orange	0.8	0.14	60	#ffa659	#ffa95d	#bf8350	#c28155	#ffa86c	0.8044315	0.144308761249727	54.0815483067796
increasing	25	attraction	Qnt3	H	amber	0.6	0.16	75	#ac7300	#bf8100	#b68947	#b78848	#c17f00	0.649669	0.154572123771914	72.8562128783697
platonic attraction	23	attraction	Attr2	B	lemon	0.9	0.16	105	#eee354	#fef365	#9b9648	#9a9649	#fcf466	0.9463	0.160927677604931	105.998845583834
emotional attraction	28	attraction	AttrG	G	chartreuse	0.7	0.14	115	#9ca831	#97a230	#90994e	#8d9a50	#93a435	0.6831315	0.135589054432078	117.748464637735
self	37	attraction	Abs2	R	lime	0.7	0.11	135	#80ae67	#89b86e	#77a061	#75a163	#86b971	0.73149	0.113139577646575	136.963536548653
romantic attraction	32	attraction	Attr1	A	emerald	0.7	0.16	150	#43b966	#4fc270	#63a471	#6ba26a	#62c063	0.727186	0.157149692274823	144.053114923865
envy / jealousy	22	attraction	Qnt3	H	viridian	0.7	0.13	165	#37b78a	#44bc90	#4ea683	#58a57b	#53bb85	0.7156375	0.125473171423396	158.234233268987
polyamory	10	attraction	Rel	Q	pine	0.65	0.09	180	#47a191	#51ad9c	#3ca694	#42a68e	#56ad97	0.686207	0.0911599015389709	174.766416770558
familial attraction	11	attraction	Attr2	B	teal	0.6	0.11	195	#009393	#1fb1b1	#31a5a5	#33a6a0	#23b1ab	0.690675	0.11269102470112	189.938405284585
tertiary attraction	9	attraction	AttrG	G	cyan	0.75	0.1	210	#55bfd1	#57b9c9	#33a3b4	#33a3b4	#57b9c9	0.732793	0.0939965609408986	209.936026999465
modest / faint	19	attraction	Abs2	R	sky	0.55	0.07	225	#407a90	#3f7e96	#419fc0	#3ea0be	#3d7f94	0.561563	0.0747344711573739	222.079871218789
slow	41	attraction	Abs1	R	sky	0.6	0.1	225	#2b8cad	#2d96b9	#419fc0	#439ec1	#2f95ba	0.629577	0.105539845765594	226.48883419564
alterous attraction	26	attraction	Attr1	A	cerulean	0.7	0.11	240	#58a7dc	#65b4ea	#549ac9	#4f9bc7	#5fb6e7	0.741757	0.110204356947417	235.901913202322
opposite / different	8	attraction	Abs3	H	azure	0.65	0.09	255	#6992c5	#688dbc	#6795cf	#6795cf	#688dbc	0.633834	0.0826653527884155	254.849153823084
decreasing	14	attraction	Qnt3	H	azure	0.6	0.14	255	#4081d2	#3a7ccc	#6795cf	#6895cf	#3a7ccc	0.581938	0.141002774414053	255.21135397594
social attraction	13	attraction	Attr2	B	violet	0.7	0.09	285	#9897d5	#9694d5	#8c89cd	#8a8acd	#9595d6	0.69365	0.0947175433900704	283.846828616532
attraction-in-general	31	attraction	AttrG	G	lavender	0.6	0.12	300	#8c6ebd	#997dcb	#9c84c7	#9885c8	#967ecc	0.647325	0.11630023187286	296.836219295817
mspec	8	attraction	AttrG	G	lavender	0.7	0.15	300	#ad87ed	#b085f4	#9c84c7	#9b84c7	#ae86f5	0.703018	0.161309924003514	299.107788953267
sexual attraction	36	attraction	Attr1	A	fuchsia	0.6	0.14	330	#ae5ea8	#b764b0	#b57baf	#b37cb2	#b465b4	0.6239095	0.144336914071581	327.362806429727
same	25	attraction	Abs3	H	mulberry	0.7	0.12	345	#d37faf	#d07cad	#be79a0	#be799f	#d17cac	0.692409	0.12015253790689	346.00374325359
love	26	attraction	Qnt3	H	mulberry	0.65	0.18	345	#d859a8	#dc5bab	#be79a0	#bf789d	#de5aa6	0.658373	0.182832235208234	347.657988659713
all / transcendent	30	gender	Qnt1	Q	rose	1	0	360	#ffffff	#ffffff	#c47890	#c5788e	#ffffff	1	0	1.14846345016628
none / agender / genderless	56	gender	Qnt2	B	amaranth	0.2	0	15	#161616	#3e3e3e	#c8787e	#c87880	#3e3e3e	0.364237	0	13.8095736509864
xenine / xenogender	51	gender	GenB	B	amaranth	0.75	0.12	15	#f08d95	#f08b94	#c8787e	#c8787a	#f18c8f	0.747753	0.122790407634573	18.3725832217812
bodily-sex-in-general	16	gender	Flu2	G	red	0.7	0.12	30	#df8071	#da7f71	#c87a6d	#c87976	#da7e7a	0.691104	0.11508356054484	22.6134692399601
questioning / confusion / dissociation	17	gender	Clar	R	vermillion	0.55	0.03	45	#816c63	#a48a7f	#c57e5e	#c67d60	#a48a80	0.654788	0.0350490965934795	42.5027706328991
strong / large	32	gender	Abs4	R	vermillion	0.6	0.16	45	#cb5a1a	#cb5e25	#c57e5e	#c38057	#c86208	0.6089155	0.154114941519709	51.9687591891671
aporine & maverine	15	gender	GenM	A	orange	0.7	0.16	60	#e58212	#fa9b44	#bf8350	#c18253	#fc9a49	0.772788	0.150515546716338	56.8595559899368
non-conformity	37	gender	Flu1	A	orange	0.75	0.13	60	#ea9952	#f4a45f	#bf8350	#bf8350	#f4a55e	0.784785	0.127671134299039	60.8394248108683
outherine	11	gender	GenC	H	amber	0.8	0.12	75	#ebb25f	#f5bd6b	#b68947	#b58947	#f4bd6a	0.832537	0.118932683563538	76.0104499452409
culturally-specific-genders	24	gender	Abs3	H	amber	0.7	0.14	75	#d0901e	#d9961e	#b68947	#b38b45	#d59917	0.722004	0.145022974173765	79.028797345418
intersex	55	gender	Sex1	Q	gold	0.85	0.18	90	#fac700	#ffd457	#aa8f44	#a59245	#fed600	0.884871	0.182053562380791	95.2649672918502
single / one	16	gender	Qnt1	Q	gold	0.6	0	90	#808080	#9e9e9e	#aa8f44	#a49245	#9e9e9e	0.699108	0	96.618026086304
placeholder	17	gender	Qnt2	B	lemon	0.8	0.15	105	#ccc23b	#e8de58	#9b9648	#a19346	#f1da53	0.882929	0.1527068857674	99.4651993898076
exobinary	25	gender	GenB	B	lemon	0.85	0.14	105	#dbd35c	#e7de67	#9b9648	#9f9447	#ecdc64	0.885819	0.141159888197829	101.892208484019
not-binary / abinary / enby	38	gender	GenB	B	lemon	0.95	0.13	105	#fbf488	#f6ee80	#9b9648	#9b9648	#f5ee81	0.933046	0.132052323601852	105.362460016255
flux	30	gender	Flu2	G	chartreuse	0.9	0.07	115	#dce4af	#dde5ae	#90994e	#909a4f	#dce6ae	0.9028985	0.0732359743180664	115.440183267512
neutral / null / neither	78	gender	Sex2	G	chartreuse	0.85	0.13	115	#cbd870	#d0dd72	#90994e	#8a9b53	#c7e078	0.8662115	0.133211723089875	120.341433283677
autonine	14	gender	Clar	R	lime	0.8	0.13	135	#99d079	#94ca76	#77a061	#7e9f5c	#9dc86f	0.781682	0.126118489360212	130.048706219013
stable / unchanging	28	gender	Flu1	A	emerald	0.65	0.08	150	#6b9d75	#76ae81	#63a471	#5ba478	#70af87	0.6993475	0.0872135840093203	155.664445228415
non-queer	35	gender	GenC	H	viridian	0.55	0.09	165	#358264	#419171	#4ea683	#4ba686	#3e9174	0.596489	0.0928240545175427	167.799424576436
multiple / many	35	gender	Qnt1	Q	pine	0.7	0.11	180	#3bb5a1	#31ab98	#3ca694	#3ea692	#34ab95	0.670689	0.108616744476186	177.723230032055
ambiguous	30	gender	Sex1	Q	pine	0.75	0.05	180	#8cb9af	#8eb7ae	#3ca694	#3ba695	#8eb7ae	0.7461025	0.0458425203334553	180.738136978617
anthrogender / trinary / quaternary	18	gender	GenB	B	teal	0.7	0.12	195	#00b5b5	#00b3b3	#31a5a5	#31a5a4	#00b3b2	0.693668	0.123241272041915	193.713306504591
nonzero	9	gender	Qnt2	B	teal	0.75	0.08	195	#6dbebe	#66b8b8	#31a5a5	#31a5a5	#66b8b7	0.731378	0.0803684879091682	194.768959952987
gender-in-general	36	gender	Flu2	G	cyan	0.7	0.09	210	#51aebd	#50b2c2	#33a3b4	#38a1b9	#53b0c7	0.710527	0.0936791655478423	216.382352095222
clear / known / certain	19	gender	Clar	R	sky	0.75	0.12	225	#45bde7	#3eb6df	#419fc0	#439ec1	#41b6e1	0.728717	0.119071125806762	226.48883419564
weak	19	gender	Abs4	R	sky	0.7	0.08	225	#64a9c4	#5f9fb7	#419fc0	#499dc5	#639dbb	0.668152	0.0747344711573739	231.898700989513
conformity / tradition	25	gender	Flu1	A	cerulean	0.6	0.05	240	#65859b	#65869e	#549ac9	#529ac9	#65869e	0.60435	0.0521223380275542	238.937990282791
masculine / male / masc-aligned	98	gender	GenM	A	cerulean	0.75	0.12	240	#5fb8f2	#5eb5ee	#549ac9	#5699ca	#61b4ef	0.741658	0.118534896682268	241.825074635534
kenochoric	35	gender	GenC	H	azure	0.55	0.08	255	#5174a0	#4d72a1	#6795cf	#6994cf	#4f72a1	0.544691	0.0846500424907684	256.281238084094
partial	30	gender	Qnt1	Q	indigo	0.8	0	270	#bebebe	#9d9d9d	#7a8fd0	#7c8fd0	#9d9d9d	0.696007	0	271.068467445242
perisex (non-intersex)	5	gender	Sex1	Q	indigo	0.6	0.1	270	#697dbc	#5d72b9	#7a8fd0	#838ccf	#676fb8	0.566647	0.112847723335138	277.367289212946
current / present	3	gender		B	violet	0.35	0.15	285	#382483	#3c2a89	#8c89cd	#8e89cd	#3e2988	0.368025	0.149524825887405	286.69744156286
binary	13	gender	GenB	B	violet	0.65	0.13	285	#8882db	#9690eb	#8c89cd	#8f88cc	#9a8fea	0.69678	0.131394807151904	287.91653253786
both / between	13	gender	Flu2	G	lavender	0.55	0.16	300	#8156c0	#9065d1	#9c84c7	#9f83c4	#9562cd	0.598947	0.162523544318148	303.928576063234
androgynous	46	gender	Sex2	G	lavender	0.65	0.17	300	#a072e6	#9e6ee5	#9c84c7	#a282c2	#a86ade	0.640396	0.174470693886117	306.705754957075
intergender	19	gender	Clar	R	plum	0.45	0.22	315	#8003a6	#7e00a3	#a97fbc	#a680bf	#7a02aa	0.443814	0.221753114714037	311.225143028373
externally influenced	6	gender	Abs4	R	plum	0.6	0.11	315	#996bad	#996cad	#a97fbc	#ac7eba	#9c6baa	0.6021515	0.107159280034573	318.126035438266
fluidity / change	45	gender	Flu1	A	fuchsia	0.7	0.13	330	#cb7fc5	#c47bbe	#b57baf	#b77bad	#c67aba	0.682275	0.125157464066099	332.648698955608
queer	55	gender	GenC	H	mulberry	0.75	0.16	345	#f381c5	#ef80c2	#be79a0	#bd79a1	#ee80c3	0.742086	0.155544523756761	344.035842851578
feminine / female / fem-aligned	95	gender	Sex1	Q	rose	0.75	0.15	360	#f982aa	#ff8bb1	#c47890	#c37893	#fd8bb6	0.772749	0.144907426897526	357.20089360778
non-human	26	other	GenB	B	amaranth	0.7	0.15	15	#ec7380	#e7717e	#c8787e	#c87978	#e87275	0.6904295	0.146157172535013	20.1523591576465
socialist	2	other		G	red	0.6	0.23	30	#ea2413	#e71604	#c87a6d	#c87a71	#e71120	0.587603	0.233611863302589	26.7913399528824
altersex	36	other	Sex2	G	red	0.85	0.11	30	#ffbaae	#f6998a	#c87a6d	#c87a71	#f6988e	0.774407	0.114911950687943	26.9365827477031
POC	22	other		R	vermillion	0.45	0.08	45	#7a462e	#7c472e	#c57e5e	#c57e5c	#7c472d	0.456342	0.0822401988586509	46.5598397464506
non-Western	15	other		A	orange	0.6	0.15	60	#bb6800	#ca7109	#bf8350	#bb864b	#c27700	0.63577	0.146780957185358	67.235269184561
vanilla	3	other	Rel	Q	gold	0.9	0.07	90	#f0dda9	#ecd8a3	#aa8f44	#b28b45	#f3d5a4	0.885343	0.0718657395135311	80.0079133432039
Christian	3	other		Q	gold	0.8	0.1	90	#d7bb70	#e6c978	#aa8f44	#aa8f44	#e7c878	0.842535	0.106056254898574	89.4020344297427
neuroqueer	25	other	GenM	A	emerald	0.8	0.12	150	#83d494	#7fcf90	#63a471	#68a36d	#85cf8c	0.787479	0.11983255639218	146.597103751785
universality	14	other	Abs3	H	viridian	0.6	0.12	165	#19966e	#279972	#4ea683	#4ba686	#1f9975	0.609748	0.115925048323524	167.672428233279
non-altersex / protosex	2	other	Sex2	G	cyan	0.5	0.07	210	#296e79	#286c78	#33a3b4	#32a3b2	#286d77	0.4948055	0.0694245045618231	208.347758042091
Jewish	10	other		H	azure	0.7	0.1	255	#73a1dc	#6a97d0	#6795cf	#6994cf	#6c96d0	0.6673025	0.0983886002738566	256.589017077213
kink & fetish	85	other	Rel	Q	indigo	0.45	0.12	270	#3c4f97	#5065af	#7a8fd0	#7690d0	#4b66af	0.52491	0.119127142782024	266.73296655693
disabled	23	other		Q	indigo	0.5	0.11	270	#4b5ea2	#5468ad	#7a8fd0	#848ccf	#5f65ac	0.534586	0.111214263164806	278.00116412004
abolitionist	7	other		R	plum	0.7	0.13	315	#bd84d6	#b47dcc	#a97fbc	#aa7fbc	#b47dcc	0.674264	0.127116543530852	315.299398500617
anarchist	6	other		A	fuchsia	0.5	0.11	330	#864a81	#7b3f77	#b57baf	#b57baf	#7b3f77	0.4642085	0.113115545523245	329.896850726201
drag	15	other	GenM	H	fuchsia	0.65	0.17	330	#c763c0	#c775c0	#b57baf	#bb7aa6	#cf72b4	0.677046	0.14062025337685	339.169424447188
feminist	6	other		Q	rose	0.6	0.14	360	#c1587d	#ce6388	#c47890	#c27897	#ca6392	0.6368005	0.140604844000473	353.124724267926

Tags by hue

hue	example-hex	colour-name	att-hex	L	C	attraction	gend-hex	L	C	gender	meta-hex	L	C	meta	other-hex	L	C	other
360	#c47890	rose	#484848	0.4	0.0	monogamy	#f982aa	0.75	0.15	feminine / female / fem-aligned	#ffffff	1.0	0.0	all / transcendent	#c1587d	0.6	0.14	feminist
15	#c8787e	amaranth	#df5f6f	0.65	0.16	aesthetic attraction	#f08d95	0.75	0.12	xenine / xenogender	#161616	0.2	0.0	none / agender / genderless	#ec7380	0.7	0.15	non-human
30	#c87a6d	red	#a43c2f	0.5	0.14	physical attraction	#ffbaae	0.85	0.11	altersex	#df8071	0.7	0.12	bodily-sex-in-general	#ea2413	0.6	0.23	socialist
45	#c57e5e	vermillion	#e57f4f	0.7	0.14	fast	#816c63	0.55	0.03	questioning / confusion / dissociation	#cb5a1a	0.6	0.16	strong / large	#7a462e	0.45	0.08	POC
60	#bf8350	orange	#ffa659	0.8	0.14	sensual attraction	#e58212	0.7	0.16	aporine & maverine	#ea9952	0.75	0.13	non-conformity	#bb6800	0.6	0.15	non-Western
75	#b68947	amber	#ac7300	0.6	0.16	increasing	#ebb25f	0.8	0.12	outherine	#d0901e	0.7	0.14	culturally-specific-genders
105	#9b9648	lemon	#eee354	0.9	0.16	platonic attraction	#fbf488	0.95	0.13	non/exo-binary	#ccc23b	0.8	0.15	placeholder
115	#90994e	chartreuse	#9ca831	0.7	0.14	emotional attraction	#cbd870	0.85	0.13	neutral / null / neither	#dce4af	0.9	0.07	flux
135	#77a061	lime	#80ae67	0.7	0.11	self	#99d079	0.8	0.13	autonine	#80ae67	0.7	0.11	self
150	#63a471	emerald	#43b966	0.7	0.16	romantic attraction	#83d494	0.8	0.12	neuroqueer	#6b9d75	0.65	0.08	stable / unchanging
165	#4ea683	viridian	#37b78a	0.7	0.13	envy / jealousy	#358264	0.55	0.09	non-queer	#19966e	0.6	0.12	universality
180	#3ca694	pine	#47a191	0.65	0.09	polyamory	#8cb9af	0.75	0.05	ambiguous	#3bb5a1	0.7	0.11	multiple / many
195	#31a5a5	teal	#009393	0.6	0.11	familial attraction	#00b5b5	0.7	0.12	anthrogender / trinary / quaternary	#6dbebe	0.75	0.08	nonzero
210	#33a3b4	cyan	#55bfd1	0.75	0.1	tertiary attraction	#296e79	0.5	0.07	non-altersex / protosex	#51aebd	0.7	0.09	gender-in-general
225	#419fc0	sky	#2b8cad	0.6	0.1	slow	#45bde7	0.75	0.12	clear / known / certain	#64a9c4	0.7	0.08	weak
240	#549ac9	cerulean	#58a7dc	0.7	0.11	alterous attraction	#5fb8f2	0.75	0.12	masculine / male / masc-aligned	#65859b	0.6	0.05	conformity / tradition
255	#6795cf	azure	#4081d2	0.6	0.14	decreasing	#5174a0	0.55	0.08	kenochoric	#6992c5	0.65	0.09	opposite / different	#73a1dc	0.7	0.1	Jewish
285	#8c89cd	violet	#9897d5	0.7	0.09	social attraction	#8882db	0.65	0.13	binary	#382483	0.35	0.15	current / present
300	#9c84c7	lavender	#8c6ebd	0.6	0.12	attraction-in-general	#a072e6	0.65	0.17	androgynous	#8156c0	0.55	0.16	both / between	#ad87ed	0.7	0.15	mspec
330	#b57baf	fuchsia	#ae5ea8	0.6	0.14	sexual attraction	#c763c0	0.65	0.17	drag	#cb7fc5	0.7	0.13	fluidity / change	#864a81	0.5	0.11	anarchist
345	#be79a0	mulberry	#d859a8	0.65	0.18	love	#f381c5	0.75	0.16	queer	#d37faf	0.7	0.12	same
90	#aa8f44	gold	#f0dda9	0.9	0.07	vanilla	#fac700	0.85	0.18	intersex	#808080	0.6	0.0	single / one	#d7bb70	0.8	0.1	Christian
270	#7a8fd0	indigo	#3c4f97	0.45	0.12	kink & fetish	#697dbc	0.6	0.1	perisex (non-intersex)	#bebebe	0.8	0.0	partial	#4b5ea2	0.5	0.11	disabled
315	#a97fbc	plum	#996bad	0.6	0.11	externally influenced	#8003a6	0.45	0.22	intergender	#996bad	0.6	0.11	externally influenced	#bd84d6	0.7	0.13	abolitionist

Results by Column (Definitions and Tagging Notes)

Column: Attraction

tag	n	colour	hex	L	C	H	definiton	notes
monogamy	2	rose	#484848	0.4	0.0	360.0	Being in one romantic/sexual relationship at a time.	Not a lot of monogamy flags. Follows the trend of conformant flags being monochromatic.
aesthetic attraction	48	amaranth	#df5f6f	0.65	0.16	15.0	Attraction based on someone’s physical appearance and/or aesthetic. Is not necissarily sexual.	I included some kink/fetish flags here⸲ like being attracted specifically to bears (an aesthetic!).
physical attraction	12	red	#a43c2f	0.5	0.14	30.0	Physical attraction. Umbrella category which includes sexual attraction⸲ sensual attraction⸲ aesthetic attraction.	I expected this to be red and it was! Lines up nicely with the red of bodily-sex-in-general.
fast	26	vermillion	#e57f4f	0.7	0.14	45.0	Attraction (or a sense of gender) which is fast: falling rapidly in love etc.	This was not a tag I anticipated existing but it makes sense – we talk about falling fast or taking it slow etc. It lives nearby passion/intense which makes sense to me.
passion / intense	27	vermillion	#d86d38	0.65	0.15	45.0	A kind of attraction or sense of gender which is intense in magnitude⸲ perhaps overpowering or overwhelming.	This came out as vermillion not so much because vermillion is used but more because the distribution is really <a bunch of reds> plus <a similar quantity of oranges>.
sensual attraction	24	orange	#ffa659	0.8	0.14	60.0	Physical attraction which is oriented around the senses: wanting to touch⸲ kiss⸲ cuddle⸲ etc. Is not necissarily sexual.	This one was consistently orange.
increasing	25	amber	#ac7300	0.6	0.16	75.0	A sense of attraction or gender which is increasing in magnitude.	This was a mix of red⸲ yellow⸲ and green.
platonic attraction	23	lemon	#eee354	0.9	0.16	105.0	Emotional attraction based around wanting to be a specific person’s friend.	Pretty high consistency here: pretty much all yellows.
emotional attraction	28	chartreuse	#9ca831	0.7	0.14	115.0	Umbrella term for attraction which is emotional (rather than physical) in nature. Includes romantic attraction⸲ platonic attraction⸲ familial attraction⸲ social attraction⸲ and alterous attraction.	This was another mix: yellow was most common followed by green then red and blue.
self	37	lime	#80ae67	0.7	0.11	135.0	Referring to the self: autosexuality⸲ self-love⸲ having a gender that is “I’m just me”⸲ etc	Teal/cyan was the most common but there was still a fair bit of variance – orange and yellow also get used.
romantic attraction	32	emerald	#43b966	0.7	0.16	150.0	Emotional attraction baseda round wanting to be in a romantic relationship with a specific person. Is not necissarily sexual.	High consistency again. Seems to be deliberately opposite (180deg) from sexual.
envy / jealousy	22	viridian	#37b78a	0.7	0.13	165.0	Feeling envy or jealousy of another person⸲ often for their gender (gender envy).	Tag added as part of post-hoc analysis to be a complement of love.
polyamory	10	pine	#47a191	0.65	0.09	180.0	Being in multiple romantic/sexual/etc relationships at once. Umbrella category which includes open relationships⸲ polyfidelity⸲ and solo polyamory.	Mostly greenish blues and bluish greens.
familial attraction	11	teal	#009393	0.6	0.11	195.0	Emotional attraction based around wanting to have a familial relationship with a specific person. This person may or may not be pre-existing family (e.g. wanting to connect with an estranged family member).	Bit of a spread here: some browns and teals but mostly cyans.
tertiary attraction	9	cyan	#55bfd1	0.75	0.1	210.0	Umbrella term for forms of attraction which are neither romantic nor sexual. Examples include sensual attraction⸲ aesthetic attraction⸲ and familial attraction.	Similar spread as familiar attraction.
modest / faint	19	sky	#407a90	0.55	0.07	225.0	A kind of attraction or sense of gender which is modest in magnitude⸲ the very opposite of passionate. A gentle sort of attraction or sense of gender which is present but may be taken for granted.	Was mostly violet/indigo but also a bunch of grey and brown.
slow	41	sky	#2b8cad	0.6	0.1	225.0	Attraction (or a sense of gender) which is taking it slow: getting to know somebody really well and openly and forming a deep connection (etc).	I included meanings like “openness in relationship communication” here. Fair bit of spread but the blue does win out.
alterous attraction	26	cerulean	#58a7dc	0.7	0.11	240.0	Emotional attraction based around wanting to be emotionally close to someone⸲ but not in a platonic or romantic way.	Had a lot of consistency.
opposite / different	8	azure	#6992c5	0.65	0.09	255.0	Referring to the opposite (or a different) gender⸲ such as in opposite-gender attraction.	Not a ton of entries here but they clearly cluster around the purpler blues.
decreasing	14	azure	#4081d2	0.6	0.14	255.0	A sense of attraction or gender which is decreasing in magnitude.	While purple was most common⸲ red and green also were contenders.
social attraction	13	violet	#9897d5	0.7	0.09	285.0	Emotional attraction based around wanting to be like a specific person. For example wanting a mentor-protégé relationship with a specific person. Or having an “intellectual crush” on someone whose work you admire.	Purple was most common but there were a handful of flags committed to a theme of social attraction being red.
attraction-in-general	31	lavender	#8c6ebd	0.6	0.12	300.0	Used for denoting attraction as a general concept.	Large spread in the entries here.
mspec	8	lavender	#ad87ed	0.7	0.15	300.0	The multisexual spectrum is an umbrella category which includes bisexuality⸲ polysexuality⸲ omnisexuality⸲ pansexuality⸲ etc.	"Fairly purple. It makes sense to me that this aligns with “both” and “androgynous"" from the gender column.\"
sexual attraction	36	fuchsia	#ae5ea8	0.6	0.14	330.0	Physical attraction which is oriented around wanting to engage in sexual activities with a specific person.	Red gets used a fair bit for this⸲ but purple was at least 2x more common. The a-spec community is pretty consistent about using purple for this.
same	25	mulberry	#d37faf	0.7	0.12	345.0	Referring to a gender being the same⸲ such as in same-gender attraction.	There’s a fair bit of overlap with the queer tag from the gender column because same-gender attraction was tagged as queer in that column. Not surprised these came out at the same hue.
love	26	mulberry	#d859a8	0.65	0.18	345.0	Deep emotional connection⸲ which can vary in many ways (e.g. romantic⸲ familial⸲ platonic).	The spread is like 1/3 purple⸲ 1/3 pink⸲ 1/3 red. It winds up aligning with queer and same (like same-gender attraction) which maybe is not a coincidence.

Column: Gender

tag	n	colour	hex	L	C	H	definiton	notes
all / transcendent	30	rose	#ffffff	1.0	0.0	360.0	Referring to all genders/attraction types at once. Also the transcendence of gender: the unlimited potential of possible genders when in a liminal space.	"White was overwhelmingly the most common colour for representing all genders - such as the white in the pangender flag which represents the blend of all genders⸲ and the white in the attractionfluid flag which represents attraction to all genders. White was also used to represent concepts such as ``the limitless void of possibilities"" (genderapathy v2)⸲ ``all genders that are accessibile within one's experience"" (fluktuaz)⸲ and transcendence (altersex). I wound up grouping these with all genders because I felt they get at the concept of everything/all - a lot of the transcendence stripes had a context of infinite possibility. White makes sense for this - in an additive colour system⸲ white is the combination of all possible colours. "
none / agender / genderless	56	amaranth	#161616	0.2	0.0	15.0	The absence of gender or attraction. Agender/genderless. Contrasted with gender neutral where there is a sense of gender and it is neutral (neither feminine nor masculine).	If I had been doing some sort of measure of mode instead of median⸲ this would have come out as black. There were a bunch of whites in here in addition to blacks. In an additive colour system⸲ black is the absence of colour. But in a subtractive colour system⸲ white is the absence of colour. I'm not too surprised that different flagmakers are referencing different colour systems. Additive colour logic tends to be most common in my data.
xenine / xenogender	51	amaranth	#f08d95	0.75	0.12	15.0	Xenogender is an umbrella term for genders based on concepts rather than sociocultural constructs. E.g. sleepygender is a gender that feels sleepy and tired. Catgender is a gender for those whose gender is feline (catboy/catgirl). The gender quality for xenogender is xenine.	This lines up with aesthetic attraction⸲ which makes sense to me as many xenogenders are aesthetic in nature. That said this was a tag which had rather little consensus: the distribution was all over the rainbow. Given the nature of xenogenders this is unsurprising.
bodily-sex-in-general	16	red	#df8071	0.7	0.12	30.0	Not a gender. Used to refer to bodily sex as an abstract concept. Contrast with gender-in-genderal.	Consistently red.
questioning / confusion / dissociation	17	vermillion	#816c63	0.55	0.03	45.0	Being uncertain of one’s gender or dissociating from the concept. I put gender dysphoria here.	Mostly greys but enough reds to make it a greyish red.
strong / large	32	vermillion	#cb5a1a	0.6	0.16	45.0	Having a sense of gender that is strong or large – more than simply having a sense of gender (nonzero).	This winds up aligning with passion/intense from the attraction column which makes sense. I thought about trying to condense those two tags into one but there just wasn’t enough internal alignment between them.
aporine & maverine	15	orange	#e58212	0.7	0.16	60.0	Maverique is a non-binary anthrogender which is independent of the gender trinary (masculine/feminine/neutral/androgyne/combinations thereof). Maveriques have a strong sense of gender and is not agender. Aporagender is an umbrella term for genders like maverique that also includes gender-neutral.	I didn’t have enough entries for aporine/aporagender and maverine/maverique to really warrant them having separate tags. They are fairly similar genders. Makes sense to me they would line up with a strong/large sense of gender and non-conformity.
non-conformity	37	orange	#ea9952	0.75	0.13	60.0	Gender expression which does not conform to societal expectations. Will be contextual: culturally-specific genders are conforming in their cultural contexts but non-conforming in other cultural contexts.	While orange was the most common choice⸲ there was not a lot of conformity as to what non-conformity should be (lol).
outherine	11	amber	#ebb25f	0.8	0.12	75.0	An umbrella term for anthrogenders that are atrinary. Includes culturally-specific genders. Includes maverique and aporagender. Does not include xenogenders.	Was a mix of browns⸲ yellows⸲ and greens.
culturally-specific-genders	24	amber	#d0901e	0.7	0.14	75.0	Culturally-specific genders such as hijra and muxe.	While amber came out as the median⸲ there are a lot of reds here. The trend tended to be to either use red or brown.
intersex	55	gold	#fac700	0.85	0.18	90.0	Intersex: one’s sex characteristics are naturally different from societal ideas of male and female sex development. Not a gender but can have a large impact on one’s sense of gender.	Intersex is not a gender. But there are a ton of intersex-exclusive gender identities⸲ modalities⸲ and expressions⸲ and they frequently use colours that have a meaning of intersex. Nearly all entries in this category were either purple or yellow (the colours of the Carpenter intersex flag). The remaining colours are white (2) and pink (1) which reference the Phlox intersex flag⸲ and one orange (1) that appears to be chosen as close to the yellow of the Carpenter flag.
single / one	16	gold	#808080	0.6	0.0	90.0	A single gender. Used for referring to single-gender attraction or the concept of having a single gender.	Not a ton of tags here⸲ about half of them are grey and then most of the rest are yellow.
placeholder	17	lemon	#ccc23b	0.8	0.15	105.0	Not a gender/attraction. Used as a placeholder in sets such as demigenders and hypergenders where this colour is swapped out for colours representing specific genders.	Mostly yellow with some oranges.
exobinary	25	lemon	#dbd35c	0.85	0.14	105.0	Umbrella term for all genders that are not connected at all to the gender binary. Does not include genders like androgynous and demifemale. Does include maverique and xenogenders. I think it excludes (most) culturally-specific genders but am not 100%.	Mostly yellow with some greens and purples.
not-binary / abinary / enby	38	lemon	#fbf488	0.95	0.13	105.0	Umbrella term for all genders that are not binary male and binary female. Includes linear combinations of male and female such as androgynous and demimale.	Like exobinary⸲ but it also has a bunch of whites added.
flux	30	chartreuse	#dce4af	0.9	0.07	115.0	The intensity with which one experiences their gender fluctuates. Their gender may or may not also change (fluidity).	A mix of orange⸲ yellow⸲ chartreuse and lime. I think the chartreuse fits (it’s a boundary colour) and it’s used in the genderfaun/floret/satyr/etc flags.
neutral / null / neither	78	chartreuse	#cbd870	0.85	0.13	115.0	Genders which are neutral: neither female nor male. Contrasted with androgynous (is female AND male). Contrasted with agender (having no gender) – here there is a sense of gender.	About ½ green⸲ ½ yellow⸲ and the rest is reds and monochromatics. Wasn’t surprised it came out on the yellower side of green. Also wasn’t surprised it came out as the opposite of androgyne⸲ as many gender neutral flags explicitly cited the use of chartreuse or green as chosen to be the opposite of androgyne purple.
autonine	14	lime	#99d079	0.8	0.13	135.0	One’s gender is unique and/or not influenced by societal concepts like male/female/intersex. Contrast with intergender (an intersex person whose gender is connected to being intersex).	Really a mix of equal amounts of yellow and teal.
stable / unchanging	28	emerald	#6b9d75	0.65	0.08	150.0	Genders which are stable/unchanging. Contrasted with flux and fluid.	About 1/3 yellow⸲ 1/3 green⸲ 1/3 blue. Kinda expected it be bluer but them’s the data.
non-queer	35	viridian	#358264	0.55	0.09	165.0	Genders which are non-queer in nature. Opposite-gender attraction. Being cisgender.	I expected this one to be blue because I associate blue with conformancy. And there were a bunch of blues⸲ but it seems like the prevailing logic was to take the opposite of queer (purple) which yields a proper green.
multiple / many	35	pine	#3bb5a1	0.7	0.11	180.0	Multiple genders but not all genders. E.g. polysexual is attracted to multiple but not all genders.	This was mostly greens but also some blues and purples. I think using a cold green makes sense to differentiate it from the warm chartreuse of gender neutrality.
ambiguous	30	pine	#8cb9af	0.75	0.05	180.0	A gender or gender expression that is ambiguous: a mix of multiple genders that does not read as one definite gender. Refers to deliberate ambiguity and not questioning/confusion.	This one was like 1/3 grey⸲ 1/3 green (usually referencing the gender trinary)⸲ and 1/3 purple (usually referencing androgyny). Green won out by a slim margin but the purples pull it up to teal.
anthrogender / trinary / quaternary	18	teal	#00b5b5	0.7	0.12	195.0	Anthrogender refers to genders which emerge from sociocultural stratification⸲ such as traditional genders like female⸲ male⸲ and culturally-specific genders like hijra and muxe. The gender trinary is an alternative view of gender to the gender binary: female⸲ male⸲ and neutral are at the three vertices⸲ and people with trinary genders are people who have genders that are some linear combination of female/male/neutral. The quaternary is a similar idea where androgyny is given a vertex.	Another mix of yellow/green/purple. I think the green is to evoke nature since anthrogender is “natural” genders. I initially tagged anthrogender and trinary separately but both had few tags and they followed the same patterns so I merged them.
nonzero	9	teal	#6dbebe	0.75	0.08	195.0	This is the opposite of agender/genderless – one has one or more genders.	Not a common one⸲ but fairly consistently a bluish green.
gender-in-general	36	cyan	#51aebd	0.7	0.09	210.0	This is not a gender but the idea of gender in the abstract. Includes being satisfied with one’s gender (whether it be questioning/known).
clear / known / certain	19	sky	#45bde7	0.75	0.12	225.0	Genders which are known with clarity/certainty. Contrasted with questioning and kenochoric.	A lot of greenish blues. It gives me the impression of clear blue water⸲ perhaps a metaphor for clarity.
weak	19	sky	#64a9c4	0.7	0.08	225.0	Having a sense of gender/attraction but it is weak in magnitude.	This was basically “cool colours” plus “dark colours”.
conformity / tradition	25	cerulean	#65859b	0.6	0.05	240.0	Gender expression that is conforming and/or rooted in tradition. Is not necissarily binary as it includes aspects of culturally-specific genders.	This was about ½ monochromatic (a common choice for conformant flags) and the rest a mix of cool colours.
masculine / male / masc-aligned	98	cerulean	#5fb8f2	0.75	0.12	240.0	Masculinity and its variants. Includes male-adjacent/masc-aligned genders such as demimale.	Like with feminine there was a strong agreement of blue for masclunie. I honestly expected this to be more cyan than azure but seems the purplier blues are more common. Like with feminine the notable exception was (once again) the Yinyang Ren flag⸲ a Chinese culturally-specific gender. The flag uses white (yang) to represent male.
kenochoric	35	azure	#5174a0	0.55	0.08	255.0	Umbrella term for genders which are vast and unknowable. Centred around the idea of kenopsia (the vibe of an abandoned place). Not the same as questioning (one has resolved their gender as unknowable). Contrast with clear/known.	Similar distribution as weak sense of gender⸲ but with dark greens added to the mix.
partial	30	indigo	#bebebe	0.8	0.0	270.0	A gender that is partial: e.g. demigirl or paraboy.	Mostly greys but some blues and purples.
perisex (non-intersex)	5	indigo	#697dbc	0.6	0.1	270.0	Refers to sex rather than gender. Perisex means not intersex.	Not a ton of tags here but they’re consistent at least.
current / present	3	violet	#382483	0.35	0.15	285.0	The current or present gender or attraction that one may have. Used for gender/attraction-fluid people to refer to a current state.	Was originally a tag in my column for values (e.g. community/joy/pride) that was moved over to this analysis to be a foil to the placeholder tag.
binary	13	violet	#8882db	0.65	0.13	285.0	Umbrella term for binary female and binary male. Also used for the abstract concept of the gender binary.	Generally if a flag is referencing binary genders it will have a pink and a blue rather than a single purple stripe⸲ but there were several that used such a stripe instead. Not surprised that purple is used as it is in between/both pink and blue.
both / between	13	lavender	#8156c0	0.55	0.16	300.0	Referring to two genders at once (usually male and female) such as in bisexuality/biromanticism.	I vacillated on separating this from androgynous and eventually landed on keeping it separate. Some of its entries are combining two genders which are not M and F⸲ or combining two non-gender states (like isogender being between cis and trans).
androgynous	46	lavender	#a072e6	0.65	0.17	300.0	Androgyne: being both feminine and masculine at the same time. Contrasted with gender-neutral which is having no gender markers.	This comes out as midway between feminine and masculine⸲ which pleases me. Notable exception to the purples: the red in the Yinyang Ren flag⸲ a Chinese culturally-exclusive androgynous gender⸲ is associated with “perfect androgyny”.
intergender	19	plum	#8003a6	0.45	0.22	315.0	An intersex person whose gender is influenced or connected to being intersex. Usually presenting in a way that is between/both feminine and masculine (in a uniquely intersex way). Opposite is extergender which I have included with autonine.	Very consistent use of intersex purple. Full disclosure: a bunch of the entries are flags that I made well before I started this project.
externally influenced	6	plum	#996bad	0.6	0.11	315.0	A gender or attraction type which is externally influenced. For example⸲ intergender: one’s gender is influenced by being intersex.	Not a lot of entries here but it seemed useful to keep around since there are genders and attractions that are fluid/flux/etc and influenced by other people.
fluidity / change	45	fuchsia	#cb7fc5	0.7	0.13	330.0	One’s gender changes over time. Gender fluidity refers to the gender changing and is distinct from flux (the intensity of one’s gender changes). I included some trans/transition concepts here.	This eventually converged on magenta but there’s a wide spread from teal to orange.
queer	55	mulberry	#f381c5	0.75	0.16	345.0	Genders which are queer in nature. Same-gender attraction. Being transgender. The queer in “queerplatonic”.	Resaonably consistently a magenta though there’s a spread from purple to green.
feminine / female / fem-aligned	95	rose	#f982aa	0.75	0.15	360.0	Femininity and its variants. Includes female-adjacent/fem-aligned genders such as demifemale.	This was overwhelmingly pink or pink-adjacent. Two notable exceptions: the black (yin) in the Yinyang-Ren flag (Chinese) represents femininity⸲ and the yellow in the Bissu flag (Indonesia) is associated with women.

Column: Other

tag	n	colour	hex	L	C	H	definiton	notes
non-human	26	amaranth	#ec7380	0.7	0.15	15.0	Not a gender. People who identify as not-quite-human⸲ perfaps for being unusual from societal ideas of what a person is (e.g. being plural⸲ being disabled in a way which makes one an actual cyborg). Includes alterhuman and otherkin.	Fair bit of overlap in this tag with xenogender⸲ which is unsurprising.
socialist	2	red	#ea2413	0.6	0.23	30.0	Indicating politics aligned with socialism and/or social democracy. Generally in a context of LGBTQ+ liberation: redistributing wealth and resources so that LGBTQ+ people can be equal members of society.	Didn’t go into this expecting socialist to show up but it’s a good reminder that pride is political! The red makes sense as it is a common socialist colour.
altersex	36	red	#ffbaae	0.85	0.11	30.0	An umbrella term for people who desire (and may transition to having) physical sex characteristics which are different from typical binary female/male arrangements.	This lines up with bodily sex⸲ which makes sense to me.
POC	22	vermillion	#7a462e	0.45	0.08	45.0	Umbrella category for people who are not white. Often used to refer specifically to black people⸲ but includes other racialized groups such as Asian and Indigenous folks.	Not surprised this came out as brown. Black and red were the next most common colours. I initally tried to keep “black people” separate from “non-white” but so many flags used POC in an ambiguous way that I gave up and merged them.
non-Western	15	orange	#bb6800	0.6	0.15	60.0	A highly ambiguous term with different meanings depending on geographic/cultural/economic/etc connotation. Generally refers to the Anglosphere (USA/Canada/Aus/NZ) plus the parts of Europe which practice Western Christianity (UK/France/Germany/etc). May or may not include countries or cultures associated with Eastern Orthodox Christianity (e.g. Russia⸲ Greece⸲ Ukraine). May or may not include Eastern Europe and/or Latin America.	Another ambiguous term that was convenient for reducing the number of distinct tags.
vanilla	3	gold	#f0dda9	0.9	0.07	90.0	The opposite of kink/fetish.	This was added post-hoc to act as a replacement for the opposite of kinky being Christian. (which is still hilarious⸲ sorry Christians.)
Christian	3	gold	#d7bb70	0.8	0.1	90.0		A lot of gold crosses and and brown to represent the wood of Jesus’ cross.
neuroqueer	25	emerald	#83d494	0.8	0.12	150.0	People whose gender feels connected to being neurodivergent. Refers more to the relationship to gender than a specific gender. Similar in concept to how intergender is when one’s gender is connected to being intersex.	The use of green I think is a good reminder that neurodiversity is greater than just autism (which tends to be associated with red and gold.)
universality	14	viridian	#19966e	0.6	0.12	165.0	What is conisdered universal or common – contrast with culturally-specific genders.	This was originally in my column for values (e.g. solidarity/community/joy) but was pulled into this analysis to be a complement to culturally-exclusive genders.
non-altersex / protosex	2	cyan	#296e79	0.5	0.07	210.0	People who are not altersex. They have (or desire to have) bodies which conform to societal ideas of female and male.	Not a lot of entries here.
Jewish	10	azure	#73a1dc	0.7	0.1	255.0		Not surprised this came out as blue⸲ which is a colour associated with Judaism.
kink & fetish	85	indigo	#3c4f97	0.45	0.12	270.0	Umbrella category for kinks & fetishes.	Given the overlap in kink/fetish communities I found it difficulty to seperate them. This is a lot of black⸲ white⸲ blue⸲ and red – these four colours get used a lot in kink/fetish flags. The blue wound up being slightly more common than the other three common colours.
disabled	23	indigo	#4b5ea2	0.5	0.11	270.0		This one I totally expected to wind up blue and nope. There were a bunch of purples!
abolitionist	7	plum	#bd84d6	0.7	0.13	315.0	Umbrella term for politics which advocate abolition of punitive⸲ carceral social institutions such as prisons⸲ the police⸲ involuntary mental health hospitalization⸲ etc.	ACAB.
anarchist	6	fuchsia	#864a81	0.5	0.11	330.0	Indicating politics aligned with anarchism. Often in a context of LGBTQ+ liberation: advocating for the abolition of state institutions which are connected to LGBTQ+ oppression.	Again⸲ pride is political. Kinda expected this one to be black given the black anarchist flag⸲ but purple gets used with the black to indicate queer anarchism specifically and that wound up more common in my data.
drag	15	fuchsia	#c763c0	0.65	0.17	330.0	People who perform drag: dressing in ways which are incongruent with their gender identity.	Was kind of relieved this didn’t land on the same pink as female because drag kings are drag too.
feminist	6	rose	#c1587d	0.6	0.14	360.0	Indicating politics aligned with feminism⸲ usually queer feminism. Often in a context of LGBTQ+ liberation.	This lines up with the tag for woman/female and that makes sense to me.

Visualization

• 
  Most common colours used for gender qualities
• 
  Distribution: gender qualities
• 
  Most common colours used for attraction types
• 
  Distribution: attraction types
• 
  Most common colours used for abstract modifiers of gender/attraction
• 
  Distribution: other identities

Discussion

While I put a lot of work into finding pride flags for this data set, a sample is at the end of the day a sample.

I was biased towards flags which have stated meanings - generally if a flag has no know meaning (and it seems like most flags don't) this meant it was not useful to my analysis. For any flag designers reading this, I encourage you to spell out your colour choices. They might be obvious to you but they probably aren't obvious to others. If you just picked colours based on vibes, that's also useful and helps autistic people like me understand the flag!

I was biased towards flags with meanings written in English. Unsurprisingly, my colour-flag associations follow Western ideas of colour - like pink and blue for female. Non-Western colour associations like the red, black, and white in the Yinyang Ren flag for androgynous, female, and male, tended to stand out as outliers.

I was biased towards flags posted on tumblr. While this is the most common place for pride flags to get posted, its demographics skew young and American. This was most obvious to me when I realized how few kink/fetish flags I had in my data set.

While I could tell from my own personal experience that things were missing when I had few kink/fetish flags, I'm sure I have other blind spots. Going searching for envy and discovering a dozen flags with gender envy as a meaning made me rather aware there are things I did not think to look for.

I'm gonna hope that if there are major types of genders/attractions/etc that I missed, people will let me know. I don't have the bandwidth to add every possible gender/attraction, so please only let me know if I'm missing a meaning that is used in at least six distinct flags (and point me to them!). Ty <3

I did not knowingly add any exclusionist flags made by TERFs/etc to my data set. I also did not put in any effort to vet the provenance of flags. My data set may very well include some. My goal was to add as many flags as I could before my hyper fixation started to run out of steam, and assume the best about flag origins.

Future work

I completed initial tagging the columns for Disability and Values, but decided to put them aside just to get everything else out the door. I plan to do another pass over those columns and incorporate the findings.

I'm also hoping to put together some recolourings of flags that match my results!