Parallel coordinate chart with Python and Matplotlib

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

from matplotlib.lines import Line2D

lemurs = pd.read_csv(
    "https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-08-24/lemur_data.csv",
    engine = "python",
    encoding = "ISO-8859-1"
)

taxonomy = pd.read_csv(
    "https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-08-24/taxonomy.csv"
)

data = (
    lemurs
    # Keep unique combinations of the following variables.
    .drop_duplicates(["dlc_id", "taxon", "sex", "age_max_live_or_dead_y"])
    # Drop observations with Not Determined sex
    .query("sex != 'ND'")
    # Compute highest age and the number of lemurs for each specie and sex
    .groupby(["taxon", "sex"])
    .agg(
        age_max=pd.NamedAgg("age_max_live_or_dead_y", max),
        count=pd.NamedAgg("age_max_live_or_dead_y", "size")
    )
    .reset_index()
)

# Pivot the data so we have two columns for each 'age_max' and 'count',
# one if for Females and the other is for Males
data = pd.pivot_table(
    data, index="taxon", columns="sex", values=["age_max", "count"]
).reset_index()

# The above produces MultiIndex column names
# Join them
data.columns = data.columns.map("_".join).str.strip("_")

# Clean one of the taxon names that is misspelled
data["taxon"] = np.where(data["taxon"] == "CMED", "CMEAD", data["taxon"])

# Merge aggregated data with the taxonomy to obtain species names
data = pd.merge(data, taxonomy, how="left", on="taxon")

data.head()

# Numbers between 0 and 1 indicate different shades of grey
GREY96 = "0.96"
GREY30 = "0.3"
GREY15 = "0.15"

BLUE = "#3A8AD8"
ORANGE = "#F2A039"

names = data["latin_name"].values
age_max_F = data["age_max_F"].values
age_max_M = data["age_max_M"].values
color =  np.where(data["age_max_F"] > data["age_max_M"], BLUE, ORANGE)

def scale(x, x_min=None, x_max=None, lower=0, upper=1):
    if x_min is None:
        x_min = np.min(x)
    if x_max is None:
        x_max = np.max(x)
    return (x - x_min) / (x_max - x_min) * (upper - lower) + lower

n_lemurs_min = min(np.min(data["count_F"]), np.min(data["count_M"]))
n_lemurs_max = max(np.max(data["count_F"]), np.max(data["count_M"]))

lower = 10
upper = 100

size_F = scale(data["count_F"].values, n_lemurs_min, n_lemurs_max, lower, upper)
size_M = scale(data["count_M"].values, n_lemurs_min, n_lemurs_max, lower, upper)

vertical_adjustments = {
    "Varecia rubra": 0,
    "Daubentonia madagascariensis": 0.6,
    "Eulemur macaco": 0.6,
    "Eulemur rubriventer": 0.4,
    "Eulemur sanfordi": 0.5,
    "Eulemur collaris": -0.1,
    "Perodicticus potto": -0.6,
    "Lemur catta": 0,
    "Eulemur flavifrons": 0,
    "Nycticebus coucang": 0,
    "Otolemur garnettii garnettii": 0,
    "Mircocebus murinus": 0,
    "Galago moholi": 0,
    "Varecia variegata variegata": 0,
    "Eulemur mongoz": 0,
    "Eulemur fulvus": 0,
    "Eulemur rufus": 0,
    "Propithecus coquereli": 0.3,
    "Cheirogaleus medius": 0.3,
    "Eulemur Eulemur": -0.1,
    "Eulemur coronatus": -0.35,
    "Eulemur albifrons": -0.25,
    "Hapalemur griseus griseus": 0,
    "Loris tardigradus": 0,
    "Mirza coquereli": 0,
    "Varecia Varecia": 0,
    "Nycticebus pygmaeus": 0
}

# Initialize plot
fig, ax = plt.subplots(figsize = (9, 11))

# Adjust figure margins, this is going to be useful later.
fig.subplots_adjust(left=0.05, right=0.90, top=0.9, bottom=0.075)

# Set background color
fig.set_facecolor(GREY96)
ax.set_facecolor(GREY96)

# Iterate over ages, colors, and sizes, adding one line and pair of dots at a time
# Note the horizontal positions are fixed at 1 and 2.
for y0, y1, c, s0, s1 in zip(age_max_F, age_max_M, color, size_F, size_M):
    ax.plot([1, 2], [y0, y1], c=c, lw=1)
    ax.scatter(1, y0, c = c, s=s0, zorder=10)
    ax.scatter(2, y1, c = c, s=s1, zorder=10)

# Space between the dot and the label
TEXT_HPADDING = 0.08

# Space between the line and the dot
LINE_HPADDING1 = 0.02
# Space between the line and the label
LINE_HPADDING2 = 0.07

for i, name in enumerate(names):
    # Take the vertical adjustment for the name
    VA = vertical_adjustments[name]
    
    # If it is odd, plot on the left
    if i % 2 != 0:
        # Add label
        x = 1 - TEXT_HPADDING
        y = age_max_F[i] + VA
        ax.text(
            x, y, name, size=11, name="Fira Sans Compressed", 
            color=GREY15, ha="right", va="center" 
        )
        
        # Add line connect dot with label
        x = [1 - LINE_HPADDING2, 1 - LINE_HPADDING1]
        y = [age_max_F[i] + VA, age_max_F[i]]
        ax.plot(x, y, color=GREY15, lw=0.5)
        
    # If it is even, plot on the right
    else:
        # Add label
        x = 2 + TEXT_HPADDING
        y = age_max_M[i] + VA
        ax.text(
            x, y, name, size=11, name="Fira Sans Compressed", 
            color=GREY15, va="center" 
        )
        # Add line connect dot with label
        x = [2 + LINE_HPADDING1, 2 + LINE_HPADDING2]
        y = [age_max_M[i], age_max_M[i] + VA]
        ax.plot(x, y, color=GREY15, lw=0.5)
fig

# Remove all spines
ax.set_frame_on(False)

# Remove horizontal and vertical ticks
ax.set_xticks([])
ax.set_yticks([])

# Expand horizontal limits to (0.5, 2.5)
ax.set_xlim(0.5, 2.5)

# Horizontal lines in the background
ax.hlines(np.arange(10, 41), 1.1, 1.9, alpha=0.2, lw=0.5, color=GREY15, zorder=0)

# Add annotations
# Note the dots added with ax.scatter are the text backgrounds
for y in np.arange(10, 45, 5):
    ax.scatter(1.5, y, s=1200, color=GREY96)
    ax.text(
        1.5, y, str(y), 
        size=22, name="Faune", color="darkgreen", weight="bold",
        alpha=0.2, va="center", ha="center"
    )


ax.text(
    0.9, 11.25, "Female", 
    name="Faune", size=27, color="darkgreen", 
    weight="bold", alpha=0.3, ha="right"
)

ax.text(
    2.1, 11.25, "Male", name="Faune", size=27, color="darkgreen", 
    weight="bold", alpha=0.3,ha="left"
)    
    
fig

# Create the values represented in the legend
labels = np.array([20, 40, 60, 80, 100])

# Scale the values to the actual dot size
sizes = scale(labels, n_lemurs_min, n_lemurs_max, lower, upper)

# Create handles
# Note the first handle is just a label, which plays the role of a title
handles = [Line2D([], [], label="Total number of lemurs hosted at DLC by species and sex", lw=0)]

# And the other handles are the dots
# See `markersize=np.sqrt(size)`. 
# This is because `s` in `ax.scatter` is measured in squared units.
handles += [
    Line2D(
        [], 
        [], 
        label=str(label), 
        lw=0, 
        color="None",
        marker="o",
        markersize=np.sqrt(size),
        markerfacecolor="black"
    )
    for label, size in zip(labels, sizes)
]

# Add legend to the figure
# bbox_to_anchor and loc="center" mean it is located in the bottom center
legend = fig.legend(
    handles=handles,
    bbox_to_anchor=[0.5, 0.05],
    handletextpad=0.5,
    handlelength=1, 
    columnspacing=1,
    loc="center", 
    ncol=6,
    frameon=False
)

# Customize style of the fonts in the legends
# The first text, which is the title, has a larger text size
for i, text in enumerate(legend.get_texts()):
    text.set_fontfamily("Fira Sans Compressed")
    text.set_color("darkgreen")
    if i == 0:
        text.set_fontsize(11)
    else:
        text.set_fontsize(9)

fig

# Note all these annotations are centered horizontally

# Title
fig.text(
    0.5, 0.94, "Longest lived lemurs at the Duke Lemur Center",
    size=28,
    name="Faune",
    weight="bold",
    ha="center",
    color="darkgreen",
)

# Sub-title
fig.text(
    0.5, 0.905, "Maximum age recorded in years (living or dead), by species and sex",
    size=18,
    name="Fira Sans Compressed",
    color=GREY30,
    ha="center",
)

# Caption
fig.text(
    0.5, 0.025, "Source: Duke Lemur Center · Graphic: Georgios Karamanis",
    size=10, 
    color=GREY30, 
    family="Fira Sans Compressed", 
    ha="center"
)


fig
# fig.savefig("plot.png", dpi=320) to store it as PNG