What is a data visualization design?

This is a summary of data visualization points published in note by Go Ando of THE GUILD, who is famous for services that focus on UX and UI.

https://note.mu/goando/n/n99f6c395ae8a

What about # 1 and # 2?

See below for # 1
- https://qiita.com/skotaro/items/cdb0732ad1ad2a4b6236
See below for # 2
- https://qiita.com/okd46/items/ee27aeee714f54a8a6c5

Notes

Regarding the font, it works on mac, but probably not on other OS, so please specify a different font.

plt.rcParams['font.family'] = 'Hiragino Sans'

It is the part of.

13. It is effective to put a graph in the table

Use pandas.

import pandas as pd
import numpy as np

%matplotlib inline
# data
apple_products = pd.DataFrame({"product":["iPhone","iPad","Mac","Services","Other"],
                              "Earnings(M dollars)":[141319,19222,25859,29980,12863],
                             "unit":[216756,43753,19251,np.nan,np.nan]})

#Value format
format_dict = {'Earnings(M dollars)':'{0:,.0f}', 'unit':'{0:,.0f}'}

#Display while setting the graph
(apple_products
 .style
 .format(format_dict)
 .hide_index()
 .bar(color="#99ceff", vmin=0, subset=['Earnings(M dollars)'], align='zero')
 .bar(color="#ff999b", vmin=0, subset=['unit'], align='zero'))

In the value format part, separate every 3 digits with a comma [^ 1]
Use the pandas styling function for tables [^ 2]
For nan, it was okay to fill it with 0, but left it as nan because services and other would not be applicable to the unit
For the .bar () part, specify align ='zero' to make the graph an appropriate size.

15. Stacked graphs are useful for comparing percentages

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


#For error avoidance in pandas
from pandas.plotting import register_matplotlib_converters
register_matplotlib_converters()

#Font settings
plt.rcParams['font.family'] = 'Hiragino Sans'  
plt.rcParams['font.weight'] = 'heavy'

#data
music_env_df = pd.DataFrame({"radio":[0.4,0.21],"CD / Download":[0.22,0.44],"Video distribution":[0.2,0.3],"Music distribution":[0.18,0.05]},
            index=["GLOBAL","JAPAN"])
#For stacked graphs
music_env_df.T.cumsum()
#Graph color
bar_colors = ["#3B7780","#98C550","#7FC2CB","#E9C645"]

#get x ticklabel
x = music_env_df.index
#Each item name in the graph
keys = music_env_df.keys()

fig,ax = plt.subplots(figsize=(7,7))

# 1.Erase the left and right frames
sides = ['left','right']
[ax.spines[side].set_visible(False) for side in sides] 

# 2.Left axis memory, memory label deleted
ax.tick_params(left=False, labelleft=False)

# 3.Change the color of the upper and lower borders
ax.spines['bottom'].set_color("dimgray")
ax.spines['top'].set_color("dimgray")

# 4.x-axis memory settings
ax.tick_params(axis='x', labelsize='x-large',color="dimgray",labelcolor="dimgray")

# 5.Plot stacked graphs and store plot information
bar_info = []
for i in range(len(keys)):
    if i == 0:
        bar_info.append(ax.bar(x, music_env_df.T.iloc[i],width=0.5,color=bar_colors[i]))
    else:
        bar_info.append(ax.bar(x, music_env_df.T.iloc[i], bottom=music_env_df.T.cumsum().iloc[i-1],width=0.5,color=bar_colors[i]))

# 6.Items of each graph
for i,one in enumerate(bar_info):
    # %Store the number of
    bar_center = [[0,0],[0,0]]
    #Stores line coordinates that emphasize differences between bar charts
    bar_line = [[0,0],[0,0]]
    for j,one_bar in enumerate(one):
        bar_center[j][0] =  one_bar.xy[0]+one_bar.get_width()/2
        bar_center[j][1] =  one_bar.xy[1]+one_bar.get_height()/2
        #Display item name
        if j == 0:    
            ax.annotate(keys[i],xy=(0,0),xycoords="data",
                       xytext=(-0.4,bar_center[j][1]),
                       ha='right',color=bar_colors[i],fontsize=16)
            bar_line[j][0] = one_bar.xy[0]+one_bar.get_width()
            bar_line[j][1] = one_bar.xy[1]
        else:
            bar_line[j][0] = one_bar.xy[0] - bar_line[0][0]
            bar_line[j][1] = one_bar.xy[1] - bar_line[0][1]
        #Display the percentage value (%)
        ax.annotate(f'{one_bar.get_height():.0%}',xy=(0,0),xycoords="data",
                   xytext=(bar_center[j][0],bar_center[j][1]),
                   ha="center",va="center",color="white",fontsize=16)
        #Show highlighted lines
        ax.arrow(bar_line[0][0],bar_line[0][1], bar_line[1][0], bar_line[1][1], head_width=0, head_length=0, ec='dimgray')
        
# 7.Set the vertical axis area
ax.set_ylim(0,1)

Create cumulative data for stacking (music_env_df.T.cumsum ())
For the # 6 emphasized line, connect the upper right of the GLOBAL bar graph to the upper left of the JAPAN bar graph.

Extra edition # 1

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


#For error avoidance in pandas
from pandas.plotting import register_matplotlib_converters
register_matplotlib_converters()

#Font settings
plt.rcParams['font.family'] = 'Hiragino Sans'  
plt.rcParams['font.weight'] = 'heavy'

#data
icing_method = pd.DataFrame([0.35, 0.19, 0.13,0.05,0.03,0.02],
                           index=['Ice bath 2 ℃', 'Ice bath 8 ℃', 'Watering+Ice massage','Watering','Fan','Vein icing'],
                           columns=['Cooling speed'])

icing_detail = ["Immerse the whole body in an ice bath at 2 ℃","Immerse the whole body in an ice bath at 8 ℃","12 ℃ watering+Ice massage",
                "Continue to apply tap water at 15 ℃ to the whole body","It hits the wind of an electric fan at room temperature of 22 ° C.","(Neck / Axilla / Inguinal)"]

#Only the color of the bar is specified as the original
ori_blue = "#71C0F9"

fig, ax = plt.subplots(figsize=(12, 6))

icing_method.plot.barh(legend=False, ax=ax, width=0.8,color=ori_blue)

# 1.Title setting
plt.title("Cooling method and cooling speed",fontsize=24,fontweight='bold',color="dimgray")

# 2.Make a lot of margin on the left
plt.subplots_adjust(left=0.35)

# 3.Reverse the order of the y-axis
ax.invert_yaxis()

# 4.Erase other than the left frame
sides = ['right', 'top', 'bottom']
[ax.spines[side].set_visible(False) for side in sides] 

# 5.y-axis x-axis tick,Erase the y-axis tick label
ax.tick_params(bottom=False, left=False,labelleft=False)

# 6.x-axis value label setting
ax.set_xticks([i*0.1 for i in range(5)])
ax.tick_params(axis='x', labelcolor="silver")

# 7.x-axis range setting (0.Up to 4 x=0.Because the grid of 4 does not come out)
ax.set_xlim(0,0.41)

# 8.x-axis grid settings
ax.grid(axis="x")

# 9.x-axis label setting
ax.set_xlabel("Body temperature drops per 10 seconds (℃)",fontsize="x-large",fontweight="bold",color="silver")

# 10.Show actual values on the right side of the bar, items and supplementary descriptions on the right side
vmax = icing_method['Cooling speed'].max()
for i, (value,main_label,sub_label) in enumerate(zip(icing_method['Cooling speed'],icing_method.index,icing_detail)):
    ax.text(value+vmax*0.02, i, f'{value:,} ℃', fontsize='x-large', va='center', color=ori_blue)
    ax.text(-0.01, i-0.1,main_label  , fontsize='xx-large', va='center',ha='right',color="dimgray")
    ax.text(-0.01,i+0.25, sub_label, fontsize='x-large' ,va='center',ha='right', color="silver")

As shown in the item and supplementary explanation, the font size cannot be changed within the same text, so they are displayed separately.