Rang is a collection of color palettes drawn from Persian, Iranian, and Persianate art and culture. Carpets, miniature paintings, tilework, manuscripts, ceramics, and architecture become practical color schemes for maps, figures, and creative work.
Note
Persian, Iranian, and Persianate
Rang draws from Persian, Iranian, and Persianate art and culture. Here, Persian refers to the Persian language and the traditions shaped through it. Iranian refers more broadly to the many cultures of Iran. The terms overlap, but they are not interchangeable. Read a note on language and scope.
The name Rang (Persian: رنگ) means "color" in Persian. I created this collection to carry the character of these artworks into data visualization, cartography, and creative coding.
Every palette traces to a documented source photograph. Its page explains the artwork, records where each color appears, and reports the nearest match found in the sampled image. It also shows pairwise color separation under four viewing simulations and includes sample plots for a practical visual check.
Use Rang with Python, R, ArcGIS Pro, QGIS, GeoLibre and HEC-RAS. The same palette can move easily from code to maps and hydraulic models.
Created by Mohsen Tahmasebi Nasab, PhD. Read the story behind Rang.
Palettes | Install | Use | How they are made | Contribute | Credits and license
Each entry pairs the artwork with its palette. Follow the link for sample plots, notes on each color and the color-vision checks.
Silk Kashan Carpet, 16th century. The Metropolitan Museum of Art, New York. Reference Persian: کاشان. Say it kah-SHAHN.
#7f3020 #ab4a47 #c07049 #c59b46 #ccac7e #e2cfb1 #8a9463 #345f72 #1a3b45
Hunting-scene tile panel at Golestan Palace, photographed 2018. Golestan Palace, UNESCO World Heritage Site. Photo by Mohsen Tahmasebi Nasab, 2018. Reference Persian: گلستان. Say it goh-leh-STAHN, Persian for rose garden.
#432f2c #ae6259 #b57f86 #b5b5ac #cbb11c #9a9a68 #45939c #577ab1 #333a80
Termeh cloth with boteh motifs, photographed 2026. Yazd textile tradition. Photo by Mohsen Tahmasebi Nasab, 2026. Reference Persian: ترمه. Say it tehr-MEH.
I made this sequential ramp for water surface elevation and depth rasters. It follows the termeh's blue ground from foam light to deep indigo. The color names come from the water they are meant to carry: Foam Mist, Glacial Blue, Powder Aqua, River Teal, Slate Blue, Deep Channel and Night Current.
#e5f0ee #c8dfe3 #a9ccd7 #7fb5c1 #5d95a8 #3e738e #274e68
Khatam panel with brass stars, 2026. Khatam marquetry tradition. Photos by Mohsen Tahmasebi Nasab, 2018 and 2026. Reference Persian: خاتم. Say it khaw-TAM.
This warm ramp comes from khatam marquetry, the Persian craft of laying wood, bone and brass into star patterns. It moves from ebony dark to brass bright. For categories, the pick order jumps between wood, metal and bone.
#1c110b #542020 #8d310e #bf480d #b27a3c #c58c51 #d9a545 #dbba94 #e5c870
Stained-glass window at Nasir al-Mulk Mosque, 2018. Nasir al-Mulk Mosque. Photograph by Tpmehdi, 2018. Reference Persian: نصیر. Say it nah-SEER, from Nasir al-Mulk Mosque in Shiraz.
This vivid spectrum comes from the stained glass of Nasir al-Mulk Mosque in Shiraz. Violet and cobalt open into cyan and green, then turn through gold to vermilion. I use it when a map or chart needs bright colors and strong contrast.
#261410 #3b1261 #4027e1 #518ffd #74ecf9 #50a877 #6dd96f #ebc05c #f04a23
Lidded enamel vessel, photographed 2026. Persian enamelwork tradition. Photo by Mohsen Tahmasebi Nasab, 2026. Reference Persian: مینا. Say it mee-NAH, Persian for enamel.
Mina is Persian for enamel, while minakari is the craft of decorating metal with it. This diverging palette comes from a lidded minakari vessel. Copper and rose meet at an enamel-white center, then turn toward pale blue, turquoise, cobalt and indigo. The blue side dominates, just as it does on the piece.
#581c1d #984946 #c19498 #e7f2f6 #9ecaee #56adee #4090a2 #193caa #07187b
Growing in thus Way, 1972, 140 x 140 cm. Iran Darroudi official website. Artwork by Iran Darroudi. Reference Persian: رستن. Say it rohs-TAN, from the Persian title Az In Gooneh Rostan.
Rostan comes from Iran Darroudi's 1972 painting Az In Gooneh Rostan, or Growing in thus Way. Near-black earth rises through wine and urgent crimson, then opens into coral, dusty rose and a pale sky. To me, the red feels both wounded and alive, while the quieter tones hold the scene in stillness.
#4b241e #603d34 #883c3e #aa3a41 #d63d3b #d26c55 #d98665 #c0947d #d2b6a1
The Wedding of Siyavush and Farangis, Folio 185v from the Shahnama of Shah Tahmasp, ca. 1525-30, painting 28.9 x 18.4 cm, page 47.3 x 32.1 cm. The Metropolitan Museum of Art, New York. Reference Persian: شاهنامه. Say it shah-nah-MEH, Persian for Book of Kings.
Shahnameh comes from the wedding of Siyavush and Farangis. Deep violet and pavilion blue give the page its quiet center, while garden green, gold, parchment and red carry the celebration around it. I wanted the palette to hold both the stillness of the couple and the music outside their room.
#41356f #4b3d97 #798cb8 #6d866a #596956 #d4b96b #d5c9b7 #b0472c #a6373e
Taste of Cherry promotional poster, 1997. Wikipedia. Promotional poster for Taste of Cherry. Reference Persian: گیلاس. Say it gee-LAAS, Persian for cherry.
Gilas takes its five colors from the promotional poster for Abbas Kiarostami's Taste of Cherry. Charcoal and muted blue hold the film's stillness, while mauve, dusty coral, and mustard yellow carry the poster's face, tree, and sunlit field. I kept the set small to match the film's spare visual language.
#58463f #598fb6 #964765 #b8715b #d8c723
Entrance iwan of the Shah Mosque, 2020. Wikimedia Commons. Photograph by Farzan95, 2020. Reference Persian: ایوان. Say it ee-VAHN, Persian for a vaulted hall open on one side.
Iwan moves from sunlit yellow into a long run of blues drawn from the tiled entrance iwan of Isfahan's Shah Mosque. Ochre and pale blue form a short threshold before turquoise, lapis, cobalt, ultramarine, indigo and midnight blue take over. I made it for maps that need a clear light-to-dark sequence with most of the visual weight carried by blue.
#efbf15 #b99241 #6897b7 #2e89ab #3263ac #0e37ae #0412cc #150c7d #06055a
The names are Persian, and they are easy once you see them spelled out. The stress lands on the last syllable.
| name | Persian | say it | meaning |
|---|---|---|---|
| Rang | رنگ | rahng, close to the English word rung | color |
| Kashan | کاشان | kah-SHAHN | a city famous for its carpets and silks |
| Golestan | گلستان | goh-leh-STAHN | rose garden, the Qajar palace in Tehran |
| Termeh | ترمه | tehr-MEH | a patterned textile associated with Yazd |
| Khatam | خاتم | khaw-TAM | marquetry of star patterns in wood, bone and brass |
| Nasir | نصیر | nah-SEER | from Nasir al-Mulk Mosque in Shiraz |
| Mina | مینا | mee-NAH | enamel, the material used in minakari |
| Rostan | رستن | rohs-TAN | to grow, from the painting's Persian title |
| Shahnameh | شاهنامه | shah-nah-MEH | Book of Kings, the wedding of Siyavush and Farangis |
| Gilas | گیلاس | gee-LAAS | cherry, the fruit |
| Iwan | ایوان | ee-VAHN | a vaulted hall open on one side |
Each palette page repeats the pronunciation beside its name.
pip install rang
The package itself has no required dependencies. The matplotlib helpers
(cmap, register and set_palette) need matplotlib, which comes with:
pip install "rang[plots]"
To install the unreleased version straight from this repository instead:
pip install "git+https://github.com/mohsennasab/Rang.git#subdirectory=python"
install.packages("remotes")
remotes::install_github("mohsennasab/Rang", subdir = "r")The ggplot2 scales need ggplot2 3.5.0 or newer. Everything else works without it.
No Rang install is needed. Grab the files from arcgis/, qgis/, geolibre/ or hecras/ and follow the matching guide, ArcGIS, QGIS, GeoLibre or HEC-RAS.
Want to try Rang without setting anything up locally? Start with the Colab-ready Python, R and hydrologic mapping examples.
A palette stores its ramp and a pick order. Ask for a few colors and Rang selects a well-separated subset. Ask for more than the palette contains and Rang interpolates along the ramp.
import rang
rang.list_palettes() # available names
rang.rang("Golestan") # the full ramp
rang.rang("Golestan", 4) # four well separated colors
rang.rang("Termeh", 100, "continuous") # smooth ramp for gridded data
# matplotlib
import matplotlib.pyplot as plt
plt.imshow(data, cmap=rang.cmap("Termeh")) # smooth
plt.imshow(data, cmap=rang.cmap("Termeh", 6)) # six fixed steps
# categorical plots pick up the palette on their own
rang.set_palette("Golestan")
# where the colors came from
rang.source("Golestan")Calling rang.register() once adds every palette to matplotlib under a
rang: name. After that any library that takes a colormap name can use
Rang, including xarray, geopandas, rioxarray and seaborn, without knowing
anything about Rang.
rang.register()
raster.plot(cmap="rang:Termeh") # xarray
gdf.plot(column="depth", cmap="rang:Iwan_r") # geopandas, reversedlibrary(Rang)
names(rang_palettes) # available names
rang("Golestan") # the full ramp, prints as a swatch
rang("Golestan", 4) # four well separated colors
rang("Termeh", 100, type = "continuous") # smooth rampWith ggplot2 there is a scale for each case, so the colors do not have to be built by hand.
library(ggplot2)
# a categorical variable
ggplot(iris, aes(Species, Petal.Length, fill = Species)) +
geom_violin() +
scale_fill_rang_d("Golestan")
# a numeric variable
ggplot(faithfuld, aes(waiting, eruptions, fill = density)) +
geom_raster() +
scale_fill_rang_c("Termeh")Every scale comes in four spellings, scale_fill_* and scale_color_* with
a scale_colour_* alias, and _d for categories against _c for numbers.
Pass direction = -1 to any of them to run the palette the other way.
Import arcgis/Rang.stylx once through the Catalog pane. Each palette appears as smooth and discrete color schemes, with its individual swatches in the color picker. The schemes work with integer and floating-point rasters. Steps are in the ArcGIS guide.
Import qgis/Rang.xml once through the Style Manager and every
palette appears in the color ramp dropdowns, smooth and discrete. The .gpl
files add the exact colors to the QGIS color picker. Steps are in the
QGIS guide.
Copy a prepared color list from geolibre/Rang.txt into a raster Custom color ramp. For vectors, use the matching graduated or categorical list to replace the generated class colors. A machine-readable bundle for notebooks and project tooling is available as geolibre/Rang.json. The GeoLibre guide covers raster layers, vector layers, legends, colorbars and Jupyter use.
Import hecras/Rang-All.xml through the RAS Mapper Surface Fill window to add the full collection, or download one palette from the hecras folder. The HEC-RAS guide walks through the interface step by step.
I use the same basic process for each palette. The complete walkthrough is in CONTRIBUTING.md.
- Colors are sampled from a photo of the artwork by k-means clustering in CIELAB, run region by region rather than over the whole image so small motifs are not averaged away by large fields of background. Sources are museum open access photos or the contributor's own photographs.
- The candidate set is trimmed to a ramp of five to twelve colors and checked against a sampled, reduced copy of the photo. Each palette page publishes the nearest CIEDE2000 distance found in that sample.
- Separation is measured under normal vision and simulated protanopia, deuteranopia and tritanopia. A pick order is computed so that the first few discrete colors stay as far apart as possible under every vision type.
- One build command regenerates the Python, R, ArcGIS, QGIS, GeoLibre and HEC-RAS files, the sample plots and the documentation, so every palette in the collection is packaged the same way.
The sample pages use real data on purpose: one day of NOAA AORC 1 km rainfall from Hurricane Harvey, a CONUS elevation grid from AWS Open Data, and, for the water palettes, a USGS flood profile and stream network from Ithaca, New York. The fetch scripts live in tools/.
New palettes are welcome. The source can come from Persian, Iranian, or Persianate art and culture. You need a photograph you have the rights to share, colors that verifiably come from that photograph, and a palette page produced by the standard tooling so it matches the rest of the collection. CONTRIBUTING.md walks through the whole process, from picking an artwork to opening a pull request, including the sample code for extracting colors from a photo and adjusting any color that does not sit right.
The two palette-making notebooks run in Google Colab and make every input and decision easy to find. Notebook 1 takes an uploaded artwork through regions, k-means, curation, adjustment, checking, and replay. Notebook 2 shows incomplete source details as warnings and lets you add or correct only the fields you choose. It then turns the verified workflow ZIP into a proposal with repository files, documentation images, software test files, and a pull request draft.
Each palette page also shows its saved extraction regions on the source photograph. The matching recipe keeps the pixel coordinates, normalized coordinates, k-means settings and color decisions together.
palettes/ one json per palette, the single source of truth
recipes/ saved regions, k-means settings and adjustment history
sources/ contributor photographs referenced by the palette files
python/ pip installable package, reads generated _palettes.py
r/ R package, reads generated palettes_data.R
arcgis/ Rang.stylx style file and the ArcGIS Pro guide
qgis/ style file, .gpl swatches and the QGIS guide
geolibre/ copy-ready and machine-readable color lists plus a usage guide
hecras/ HEC-RAS custom color-ramp imports and the interface guide
tools/ complete Colab workflow notebooks and command-line tools
docs/ one page per palette with regions, previews and sample plots
data/ small rainfall, elevation and flood-depth grids used by the examples
examples/ Colab-ready palette and hydrologic mapping notebooks
Rang source code is licensed under the MIT License. To the extent that copyright or database rights apply, the palette names, color values, descriptions, and ordering are dedicated to the public domain under CC0 1.0 Universal. The palettes may therefore be used for any purpose without permission or credit, though a link back is always appreciated. The complete file-by-file scope is in the licensing guide.
Artwork photographs are not covered by the MIT License or CC0 unless stated otherwise. The Golestan, Termeh, Khatam and Mina source photographs remain copyright Mohsen Tahmasebi Nasab, with all rights reserved. The Kashan palette draws on the Silk Kashan Carpet at The Metropolitan Museum of Art, whose qualifying Open Access images are available under CC0. Shahnameh also uses a public-domain image from The Met's Open Access collection. Its folio shows the wedding of Siyavush and Farangis. The Golestan palette comes from a photograph of the tilework at Golestan Palace in Tehran, taken by Mohsen Tahmasebi Nasab in 2018. The Termeh palette comes from his photograph of a termeh cloth, 2026. The Smithsonian's termeh record documents the Yazd tradition and the boteh motif. The Khatam palette comes from his photographs of khatam marquetry, 2026, and of a handicraft shop in the Isfahan bazaar, 2018. The Encyclopaedia Iranica article on crafts in Isfahan documents the craft's history in Shiraz, Isfahan and Tehran. The Nasir palette draws on a stained-glass photograph by Tpmehdi, 2018, and an interior photograph by MohammadReza Domiri Ganji, 2013. Both are licensed under CC BY-SA 4.0. The Mina palette comes from Mohsen Tahmasebi Nasab's photograph of a lidded enamel vessel, taken in 2026. Its setting photograph, Iranian vitreous enamel, cropped, was made by Wikimedia Commons user مانفی in 2012 and cropped by Joalbertine in 2020. It is licensed under CC BY-SA 3.0.
.JPG){kind=link}
The Rostan palette comes from Iran Darroudi's 1972 painting Growing in thus Way. The low-resolution reference image is all rights reserved and is included only to identify and discuss the source work.
The Gilas palette comes from the promotional poster for Abbas Kiarostami's 1997 film Taste of Cherry. The Wikipedia file page identifies the poster as copyrighted and non-free. The 220-pixel reference is included only to identify and discuss the poster and the palette drawn from it.
The Iwan palette uses two photographs of the Shah Mosque in Isfahan. The extraction photograph is by Farzan95, 2020. The page photograph is by Diego Delso, 2016. Both are licensed under CC BY-SA 4.0.
The sample plots use public data: rainfall from the NOAA Analysis of Record for Calibration, elevation from the AWS Open Data terrain tiles built on SRTM, GMTED2010 and ETOPO1, water surface elevation from the USGS flood inundation study of the Ithaca creeks, and stream networks from the USGS Network Linked Data Index, with NHDPlusV2 stream order from USGS Fabric. Detailed source status, requested credits, and data notices are recorded in the third-party notices.
The package interface tries to follow the conventions set by MetBrewer and wesanderson.