Royal Family Popularity Ranker¶

Analysis by Jason Muteham

IMAGE

UK ONS Live Birth Dataset 1996-2021¶

Popularity is based on the number of babies named after royalty.

To compare names against the Office for National Statistics UK baby name dataset, alter the code in the cell below.

females = ["girl name","another girl"]

males = ["boy name","another boy name","another boy"]

Charts will be plotted using Bokeh, Matplotlib, Seaborn & Plotly

Using the ONS live birth data 1996-2021 www.ons.gov.uk

Photo by Markus Spiske on Unsplash

Setup names to compare¶

In [ ]:
#Setup names for family.
females = ["Meghan","Catherine"]
males = ["Harry","William"]
family = females + males

IMPORTS¶

In [ ]:
import pandas as pd

import matplotlib.pyplot as plt
%matplotlib inline

from bokeh.plotting import figure, show
from bokeh.models import ColumnDataSource
from bokeh.models import NumeralTickFormatter
from bokeh.io import output_notebook
output_notebook()

import plotly.express as px
import plotly.io as pio 
pio.renderers.default = "vscode+notebook"

import seaborn as sns
Loading BokehJS ...

Create lists for later data wrangling¶

In [ ]:
#Create some lists to select or reformat data later
count_cols = []     # "1996 Count", ... "2021 Count" ** Used to drop/rename columns
rank_cols = []      # "1996 Rank", ... "2021 Rank" ** Used to drop/rename columns
years = []          # "1996", ... "2021" ** Used to wrangle data and rename columns
colours = ["red","blue","yellow","orange","brown","cyan","black"] # for charts
for f in range(1996,2022):
    count_cols.append(str(f) + ' Count')
    rank_cols.append(str(f) + ' Rank')
    years.append(str(f))
In [ ]:
#Import ONS data
df1 = pd.read_excel("babynames1996to2021.xlsx","1",skiprows =7)     # Boys names in workbook 1
df2 = pd.read_excel( 'babynames1996to2021.xlsx',"2",skiprows =7)    # Girls names in workbook 2
df1 = df1.set_index("Name")
df2 = df2.set_index("Name")
In [ ]:
#Replace NaN with 0
df1 = df1.replace("[x]",0)
df2 = df2.replace("[x]",0)

Filter our names to new DataFrame¶

In [ ]:
#Filter names from ONS data
Girls = df2.loc[females]
Boys = df1.loc[males]
df_family = pd.concat([Girls, Boys], axis=0)
In [ ]:
#Drop rank and count columns
df_family_counts = df_family.drop(columns=rank_cols)
df_family_ranks = df_family.drop(columns=count_cols)
In [ ]:
#Rename columns "1996 Count" -> "1996" and "1996 Rank" -> "1996"
df_family_counts.columns = df_family_counts.columns.str.replace(" Count","")
df_family_ranks.columns = df_family_ranks.columns.str.replace(" Rank","")
In [ ]:
#Pivot DFs so rows are date observations
df_family_counts = pd.pivot_table(df_family_counts, values = years, columns=["Name"])
df_family_counts.index.name = "Year"
df_family_counts = df_family_counts.reindex(columns=family)

df_family_ranks = pd.pivot_table(df_family_ranks, values = years, columns=["Name"])
df_family_ranks.index.name = "Year"
df_family_ranks = df_family_ranks.reindex(columns=family)

Tables¶

UK Ranking¶

In [ ]:
#Display the rank table
df_family_ranks.sort_index(ascending=False).head()
Out[ ]:
Name Meghan Catherine Harry William
Year
2021 1286 457 7 21
2020 710 416 8 20
2019 641 417 5 17
2018 431 419 3 14
2017 797 313 2 11

Number of Top 20 entries¶

In [ ]:
#Display the rank table
for name in family:
    df_temp = df_family_ranks[name]
    print(name + (' ' * (10 - len(name))) +' \t' + str(df_temp[df_temp <= 20].count()))

Meghan     	0
Catherine  	0
Harry      	26
William    	24

Number of Top 10 entries¶

In [ ]:
#Display the rank table
for name in family:
    df_temp = df_family_ranks[name]
    print(name + (' ' * (10 - len(name))) +' \t' + str(df_temp[df_temp <= 10].count()))

Meghan     	0
Catherine  	0
Harry      	19
William    	15

Number of No 1's¶

In [ ]:
#Display the rank table
for name in family:
    df_temp = df_family_ranks[name]
    print(name + (' ' * (10 - len(name))) +' \t' + str(df_temp[df_temp == 1].count()))

Meghan     	0
Catherine  	0
Harry      	2
William    	0

Highest position in charts¶

In [ ]:
#Display the rank table

df_fs = pd.pivot_table(df_family_ranks,columns="Year")
df_fs = df_fs.unstack()
df_fs = df_fs.reset_index()
df_top_rank = pd.DataFrame()
df_low_rank = pd.DataFrame()

for name in family:
    df_tmp = df_fs[df_fs["Name"]==name]
    df_tmp.columns=["Year","Name","Rank"]
    df_tmphigh = (df_tmp[df_tmp["Rank"].min()==df_tmp["Rank"]])
    df_tmplow = (df_tmp[df_tmp["Rank"].max()==df_tmp["Rank"]])
    df_top_rank = pd.concat([df_top_rank,df_tmphigh])
    df_low_rank = pd.concat([df_low_rank,df_tmplow])
    
df_top_rank = df_top_rank.reindex(columns=["Name","Year","Rank"])
df_top_rank = df_top_rank.set_index("Name")
df_top_rank.sort_values(by=["Rank","Year"],ascending = [True,False])
Out[ ]:
Year Rank
Name
Harry 2012 1
Harry 2011 1
William 2010 7
William 2006 7
Catherine 1996 68
Meghan 1998 237

Lowest position in charts¶

In [ ]:
df_low_rank = df_low_rank.reindex(columns=["Name","Year","Rank"])
df_low_rank = df_low_rank.set_index("Name")
df_low_rank.sort_values(by=["Rank","Year"])
Out[ ]:
Year Rank
Name
Harry 1996 17
Harry 1997 17
William 1996 21
William 2021 21
Catherine 2021 457
Meghan 2021 1286

Number of UK Babies¶

In [ ]:
df_family_counts.sort_index(ascending=False).head()
Out[ ]:
Name Meghan Catherine Harry William
Year
2021 25 95 3089 2093
2020 52 102 3209 2215
2019 61 101 3823 2667
2018 101 104 4512 3015
2017 49 155 5031 3437
In [ ]:
#Create totals DF
df1_totals = df1.drop(columns=rank_cols).sum()
df2_totals = df2.drop(columns=rank_cols).sum()
df1_totals = df1_totals.reset_index()
df2_totals = df2_totals.reset_index()
df_totals = pd.concat([df1_totals, df2_totals], axis=0)
df_totals = df_totals.replace(count_cols,years)
df_totals.columns=["Years","Total"]
#df_totals["date"]=pd.to_datetime(df_totals["Years"])
df_totals = df_totals.groupby("Years").sum()
#df_totals
In [ ]:
#Calculate accumulative percentage change
df_family_diff = pd.DataFrame(df_family_counts,copy=True)
temp_col = [] # Build a list of temp columns to drop later
for name in family:
 #   df_family_diff[name + " Diff"] = df_family_diff[name].diff()
 #   temp_col.append(name + " Diff")    
    df_family_diff[name + " %"] = df_family_diff[name].pct_change()
df_family_diff = df_family_diff.fillna(0)

for name in family:
    df_family_diff[name + " %"] = df_family_diff[name + " %"].cumsum()

# Drop the columns not required for plotting
temp_col += family
df_family_diff = df_family_diff.drop(columns=temp_col)

Bokeh Charts¶

In [ ]:
#Chart dimensions
chart_w = 1080
chart_h = 545
In [ ]:
source = ColumnDataSource(df_totals)
p = figure(title="Total UK baby names", x_axis_label="Year", y_axis_label="Babies", x_range=years,width=chart_w, height=chart_h)
p.line(source=source, x="Years",y="Total")
show(p)
In [ ]:
# create a new plot with a title and axis labels
source = ColumnDataSource(df_family_counts)
p = figure(title="UK baby count for selected names", x_axis_label="", y_axis_label="Babies",x_range=years, width=chart_w, height=chart_h)

for index,name in enumerate(family): 
    p.line(x = "Year", y = name, legend_label=name, line_color=colours[index],line_width=2,source=source)

show(p)
In [ ]:
# create a new plot with a title and axis labels
source = ColumnDataSource(df_family_ranks)
p = figure(title="UK baby name rankings", x_axis_label="Year",y_range=(1500,1), y_axis_label="Name UK rank",x_range=years, width=chart_w, height=chart_h)

for index,name in enumerate(family): 
    p.line(x = "Year", y = name, legend_label=name, line_color=colours[index],line_width=2,source=source)

show(p)
In [ ]:
# create a new plot with a title and axis labels
source = ColumnDataSource(df_family_diff)
p = figure(title="Name Popularity % Change", x_axis_label="Year", y_axis_label="%",x_range=years, width=chart_w, height=chart_h)

for index,name in enumerate(family): 
    p.line(x = "Year", y = (name + " %"), legend_label=name, line_color=colours[index],line_width=2,source=source)

show(p)

Matplotlib Charts¶

In [ ]:
df_totals.plot(kind="line",title="Total UK baby names",figsize=(15,6),grid=True)
plt.show()
In [ ]:
df_family_counts.plot(title="UK baby count for selected names",figsize=(15,6),grid=True)
Out[ ]:
<AxesSubplot: title={'center': 'UK baby count for selected names'}, xlabel='Year'>
In [ ]:
df_family_ranks.plot(title="UK baby name rankings",figsize=(15,6),grid=True)
plt.ylim(1500,1)
Out[ ]:
(1500.0, 1.0)
In [ ]:
df_family_diff.plot(title="Name Popularity % Change",figsize=(15,6),grid=True)
Out[ ]:
<AxesSubplot: title={'center': 'Name Popularity % Change'}, xlabel='Year'>

Seaborn Charts¶

In [ ]:
fig, ax = plt.subplots(figsize=(15, 6))
sns.lineplot(df_totals).set(title="Total UK baby names")
plt.grid(True)
In [ ]:
fig, ax = plt.subplots(figsize=(15, 6))
sns.lineplot(df_family_counts).set(title="UK baby count for selected names")
plt.grid(True)
In [ ]:
fig, ax = plt.subplots(figsize=(15,6))
sns.lineplot(df_family_ranks).set(title="UK baby rank for selected names")
plt.ylim(1500,1)
plt.grid(True)
In [ ]:
fig, ax = plt.subplots(figsize=(15,6))
sns.lineplot(df_family_diff).set(title="Name Popularity % Change")
plt.grid(True)

Plotly Charts¶

In [ ]:
px.line(df_totals, title = "Total UK baby names")
In [ ]:
px.line(df_family_counts, title="UK baby count for selected names")
In [ ]:
fig = px.line(df_family_ranks, title="UK baby rank for selected names")
fig.update_yaxes(autorange="reversed")
In [ ]:
px.line(df_family_diff, title="Name Popularity % Change")

Conclusion¶

Harry and William are popular royal names rarely out of the top 20. Harry is the only name to have reached the top spot, not once but twice. William on the other hand has only reached number 7 and is currently out of the top 20 at position 21.

Catherine and Meghan are not popular names, they have never been in the top 20. Both Catherine amd Meghan have been in slow decline since the late 90's. Catherine received a small increase during the year of her royal wedding. Meghan had a sharp increase during the year of her royal wedding.

2021 see all royal names at their lowest rankings, with the exception of Harry who still has a top 10 position at number 7.

Limitation of investigation¶

The obvious limitation is that of alternative name spelling and aliases. Meghan has many different spellings in the ONS dataset (Megan, Meggan, Megann, etc). Catherine is also commonly know as Kate. The code currently doesn't allow for these aliases.

From the start the code was designed to be parameter driven and to avoid any hardcoding, this would allow the codebase to be reused for similar name analysis. Adding parameters to allow for alias checking is a level of complexity that maybe added in the future.