Before we begin, we will change a few settings to make the notebook look a bit prettier:

In [1]:

%%html
<style> body {font-family: "Calibri", cursive, sans-serif;} </style>

Visualization of the
Favorite Pokémon Survey Results

In this notebook, I will generate different visualizations of the Favorite Pokémon Survey results. The data was collected and made public by reddit user mamamia1001. Full credit goes to him/her. I will also be including a few comments on the results and how I interpret them. If you want a more simplified version (i.e., only the plots, no code), you can find it here. For more info, take a look at the README file.

Alright, let's get started.

Preliminaries¶

First, let's import all the relevant packages, configure some plotting options, and define some basic (path) variables.

In [2]:

%matplotlib inline

import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib as mpl
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
from mpl_toolkits.axes_grid1.inset_locator import inset_axes
import pathlib
from PIL import Image
import requests
from io import BytesIO

import pokefunctions

In [3]:

mpl.rcParams['font.sans-serif'] = 'Calibri'
mpl.rcParams['font.family'] = 'sans-serif'
sns.set(font_scale=1.75)
sns.set(font = 'Calibri')
sns.set_style('ticks')

plt.rc('axes.spines', top=False, right=False)

# Control font sizes.
plt.rc('font', size=12) # Default text size
plt.rc('axes', titlesize=16) # Axes title
plt.rc('axes', labelsize=16) # x and y labels
plt.rc('xtick', labelsize=12) # x tick labels
plt.rc('ytick', labelsize=12) # y tick labels
plt.rc('legend', fontsize=12) # Legend
plt.rc('figure', titlesize=16) # Figure title

In [4]:

# Define paths.
PATH_DATA = pathlib.Path(r'../data')
PATH_IMAGES = pathlib.Path(r'../images')

Read data¶

I downloaded a copy of the original results as an Excel file. Let's read just the relevant parts:

In [5]:

df = pokefunctions.read_raw_data(PATH_DATA/'responses.xlsx')
df.head(3)

Out[5]:

	name	votes	types	generation	family
1	Bulbasaur	710.0	Grass Poison	1	Bulbasaur
2	Ivysaur	83.0	Grass Poison	1	Bulbasaur
3	Venusaur	127.0	Grass Poison	1	Bulbasaur

Original Analysis¶

As a sanity check, let's (re)compute the already reported numbers.

Number of valid votes¶

Original result: 52725

From the aggregated results:

In [6]:

n_votes = df['votes'].sum()
print(n_votes)

52725.0

Average votes per Pokémon¶

Original result: 65

In [7]:

n_pokemon = len(df['name'].unique())
print(n_votes/n_pokemon)

65.17305315203956

Average votes per Pokémon per generation¶

In [8]:

df_votes_generation = df.groupby(['generation']).mean()
print(df_votes_generation)

                 votes
generation            
1           121.119205
2            91.150000
3            70.140741
4            70.383178
5            28.147436
6            29.916667
7            20.181818

In [9]:

fig, ax = plt.subplots(1, 1, figsize=[5, 5])
ax.plot(df_votes_generation.index, df.groupby(['generation']).mean()['votes'], linewidth=5)
ax.set_xticks(df_votes_generation.index)
ax.set_ylim((0, 130))
ax.set_title("Average votes per generation", fontdict = {'weight':'bold'})
ax.set_xlabel("Generation")
ax.set_ylabel("Average votes")
fig.savefig(PATH_IMAGES/'generation_average.png', dpi=1000, bbox_inches='tight')
plt.show();

Generation colors were taken from Bulbapedia.

In [10]:

generation_palette = pokefunctions.generation_palette()

fig, ax = plt.subplots(1, 1, figsize=[7, 7])
sns.boxplot(x='generation', y='votes', data=df, ax=ax, palette=generation_palette)
ax.set_title("Vote dispersion per generation", fontdict = {'weight':'bold'})
ax.set_xlabel("Generation")
ax.set_ylabel("Votes")
fig.savefig(PATH_IMAGES/'generation_boxplot.png', dpi=1000, bbox_inches='tight')
plt.show();

The outliers show us Pokémon with a high number of votes. As we will see in a second, it is clear that no Pokémon from generations 5, 6, or 7 cracks the top preference spots.

Pokémon ranking¶

Sort Pokémon from most to least popular:

In [11]:

df_sorted = df.sort_values('votes', ascending=False)

The top 10 (most popular) Pokémon:

In [12]:

df_sorted.head(10)

Out[12]:

	name	votes	types	generation	family
6	Charizard	1107.0	Fire Flying Dragon	1	Charmander
94	Gengar	1056.0	Ghost Poison	1	Gastly
59	Arcanine	923.0	Fire	1	Growlithe
1	Bulbasaur	710.0	Grass Poison	1	Bulbasaur
257	Blaziken	613.0	Fire Fighting	3	Torchic
197	Umbreon	607.0	Dark	2	Eevee
448	Lucario	604.0	Fighting Steel	4	Riolu
282	Gardevoir	585.0	Psychic Fairy	3	Ralts
133	Eevee	581.0	Normal	1	Eevee
149	Dragonite	551.0	Dragon Flying	1	Dratini

In [13]:

fig, ax = plt.subplots(1, 1, figsize=[7, 7])
sns.barplot(x=df_sorted.head(10)['votes'], y=df_sorted.head(10)['name'], hue='generation', palette=generation_palette, dodge=False, data=df_sorted.head(10), ax=ax)
pokemon_image = pokefunctions.get_sprite(df_sorted.index[0])
if pokemon_image != None:
    ax_ins = inset_axes(ax, width=2, height=2, loc='lower right')
    ax_ins.imshow(pokemon_image)
    ax_ins.axis('off')
ax.set_title("Most popular Pokemon", fontdict={'weight':'bold'})
ax.set_xlabel("Votes")
ax.set_ylabel("")
ax.yaxis.set_tick_params(labelsize=16)
ax.legend(frameon=False, loc='upper right', bbox_to_anchor=(1.25, 1)).set_title("Generation")
fig.savefig(PATH_IMAGES/'popular.png', dpi=1000, bbox_inches='tight')
plt.show();

Now we know who were those outliers from earlier ;) .

Pokémon ranking by family¶

The top 10 (most popular) Pokémon families:

In [14]:

df_families = df[['votes', 'family']].groupby(['family']).sum().sort_values('votes', ascending=False).head(10)
df_families

Out[14]:

	votes
family
Eevee	2865.0
Charmander	1551.0
Gastly	1384.0
Growlithe	1199.0
Cyndaquil	1092.0
Squirtle	1066.0
Ralts	924.0
Bulbasaur	920.0
Scyther	818.0
Mudkip	806.0

In [15]:

fig, ax = plt.subplots(1, 1, figsize=[7, 7])
sns.barplot(x=df_families['votes'], y=df_families.index, data=df_families, ax=ax)
ax.set_title("Most popular Pokemon families", fontdict={'weight':'bold'})
ax.set_xlabel("Votes")
ax.set_ylabel("Pokemon Family")
ax.yaxis.set_tick_params(labelsize=16)
fig.savefig(PATH_IMAGES/'popular_family.png', dpi=1000, bbox_inches='tight')
plt.show();

Pokémon ranking by generation¶

The most popular Pokémon of each generation:

In [16]:

generations = list(df['generation'].unique().astype(int))
generations.insert(0, 'all')

for generation in generations:
    if generation == 'all':
        idx = df.groupby('generation')['votes'].transform(max) == df['votes']
        df_generation = df[idx]
        title_str = "each generation"
        file_name = "all_generations"

    else:
        df_temp = df.query('generation == "' + str(generation) + '"')
        df_generation = df_temp.sort_values('votes', ascending=False).head(10)
        title_str = "generation " + str(generation)
        file_name = "generation" + str(generation)
    
    fig, ax = plt.subplots(1, 1, figsize=[7, 7])
    sns.barplot(x=df_generation['votes'], y=df_generation['name'], hue='generation', palette=generation_palette, dodge=False, data=df_generation, ax=ax)
    if generation == 'all':
        ax.legend(frameon=False, loc='lower right').set_title("Generation")
    else:
        # Remove legend.
        ax.get_legend().remove()
        
        # Add sprite of the most popular Pokemon of each type.
        pokemon_image = pokefunctions.get_sprite(df_generation.index[0])
        if pokemon_image != None:
            ax_ins = inset_axes(ax, width=2, height=2, loc='lower right')
            ax_ins.imshow(pokemon_image)
            ax_ins.axis('off')
    
    ax.set_title("Most popular Pokemon of " + title_str, fontdict={'weight':'bold'})
    ax.set_xlabel("Votes")
    ax.set_ylabel("")
    ax.yaxis.set_tick_params(labelsize=16)
    fig.savefig((PATH_IMAGES/(file_name + '.png')), dpi=1000, bbox_inches='tight')
    plt.show()
pass; # Supress plot output.

Visualization of the Favorite Pokémon Survey Results