How to create a color dict for multiple items from a data frame?

Question

I have over 40 or so unique items in a dataframe column (e.g. 'a','b','c'......). I want to create a colour dictionary without having to type something like the following. So that each unique item has a unique colour - How can i create a dictionary like below which would enable me to give a unique colour to each item.

Any help would be greatly appreciated.

    color_discrete_map = {'a':'rgb(42,9,4)', 
                              'b':'rgb(111,203,209)',
                              'c':'rgb(55,165,172),',
                              'd':'rgb(29,127,136)',
                              'e':'rgb(2,84,92)',
                              'f':'rgb(4,37,42)',
#....
}

Answer 1

You can use seaborn.color_palette:

import seaborn as sns
sns.color_palette("husl", 10) # 10 colors

The returned object contains tuples with the RGB values:

>>> list(sns.color_palette("husl", 10))
[(0.9677975592919913, 0.44127456009157356, 0.5358103155058701),
 (0.8616090647292522, 0.536495730113334, 0.19548899031476086),
 (0.6804189127793346, 0.6151497514677574, 0.19405452111445337),
 (0.46810256823426105, 0.6699492535792404, 0.1928958739904499),
 (0.20125317221201128, 0.6907920815379025, 0.47966761189275336),
 (0.21044753832183283, 0.6773105080456748, 0.6433941168468681),
 (0.2197995660828324, 0.6625157876850336, 0.7732093159317209),
 (0.433280341176423, 0.6065273407962815, 0.9585467098271748),
 (0.8004936186423958, 0.47703363533737203, 0.9579547196007522),
 (0.962272393509669, 0.3976451968965351, 0.8008274363432775)]

NB. 40 colors is a lot for the human eye to be able to discriminate all, but have a look at the options of the method, there are may possibilities

other examples.

"Set2":

"Paired":

Finally, an example of how to construct a dictionary:

dict(zip(list('ABCDEFG'), [tuple(int(c*255) for c in cs) for cs in sns.color_palette("husl", 7)]))

output:

{'A': (246, 112, 136),
 'B': (197, 147, 49),
 'C': (130, 168, 49),
 'D': (52, 175, 137),
 'E': (55, 170, 186),
 'F': (128, 150, 244),
 'G': (244, 93, 235)}

Answer 2

Simple enough, just grab a list of colours then join along the axis.

e.g

df_colors = pd.read_csv('https://raw.githubusercontent.com/codebrainz/color-names/master/output/colors.csv',header=None)

df_colors.columns = ['color_name', 'color_description', 'hex_value', 'r', 'g','b']

---

print(df_colors)

          color_name      color_description hex_value    r    g    b
0    air_force_blue_raf   Air Force Blue (Raf)   #5d8aa8   93  138  168
1   air_force_blue_usaf  Air Force Blue (Usaf)   #00308f    0   48  143
2  air_superiority_blue   Air Superiority Blue   #72a0c1  114  160  193
3       alabama_crimson        Alabama Crimson   #a32638  163   38   56
4            alice_blue             Alice Blue   #f0f8ff  240  248  255

---

assuming your dataframe has unique values then you can just join along the indices.

print(df)

  name
0    a
1    b
2    c
3    d

df1 = df.join(df_colors[['r','g','b']].astype(str).agg(','.join,axis=1).to_frame('rgb'),how='left')

print(df1)

  name          rgb
0    a   93,138,168
1    b     0,48,143
2    c  114,160,193
3    d    163,38,56

---

if you specifically want a dictionary, you can zip then after creating two iterables.

n = df['name'].unique()

c = df_colors[['r','g','b']].astype(str).agg(','.join,axis=1).tolist()

dict(zip(n,c))

{'a': '93,138,168', 'b': '0,48,143', 'c': '114,160,193', 'd': '163,38,56'}