Emigration from Denmark Last Year¶
Revisiting an old post, looking into foreigners leaving Denmark
Who is leaving and where are they going?
In [1]:
# Importing my libraries
import pandas as pd
import folium
import numpy as np
import matplotlib.pyplot as plt
import squarify
In [2]:
# Data from UDVAN on statbank.dk
df_age = pd.read_csv('ages2024.csv', header=None, encoding='ISO-8859-1')
df_age.columns = ['Citizenship', 'Age Bracket', 'Emigrant Count']
df_age = df_age.drop('Citizenship', axis=1)
# Ensure 'Emigrant Count' is numeric, and coerce any non-numeric values to NaN
df_age['Emigrant Count'] = pd.to_numeric(df_age['Emigrant Count'], errors='coerce')
# Drop rows with a 0 or NaN 'Emigrant Count'
df_age = df_age[df_age['Emigrant Count'].notna() & (df_age['Emigrant Count'] != 0)]
categories = df_age['Age Bracket']
values = df_age['Emigrant Count']
# Create the figure and axis
fig, ax = plt.subplots(figsize=(10, 6))
colors = ['#0093B7']
# Create a horizontal bar chart
ax.barh(categories, values, color=colors)
# Add labels and title
ax.set_xlabel('Number of People')
ax.set_title('Emigrant Distribution by Age Group in 2024')
ax.xaxis.set_visible(True)
ax.tick_params("x", rotation=45)
# Add values on the bars for clarity
#for i, v in enumerate(values):
# ax.text(v + 1000, i, str(v), color='black', va='center')
# Remove the borders (spines)
#for spine in ax.spines.values():
# spine.set_visible(False)
# Save the chart as a PNG file
plt.savefig('foreign_emigration_distribution_2024.png', bbox_inches='tight')
# Show the chart
plt.tight_layout()
plt.show()
In [3]:
# Group data into the requested age ranges
grouped_data = {"0-20 years": 0, "21-39 years": 0, "40-69 years": 0, "70+ years": 0}
for index, row in df_age.iterrows():
age_range = row['Age Bracket']
count = row['Emigrant Count']
# Extract the lower bound of the age range
if age_range == "110 years +":
lower_bound = 110
else:
lower_bound = int(age_range.split("-")[0])
# Categorize based on age groups
if lower_bound <= 19: # 0-19 covers the 0-20 years range
grouped_data["0-20 years"] += count
elif lower_bound <= 39: # 20-39 covers the 21-39 years range
grouped_data["21-39 years"] += count
elif lower_bound <= 69: # 40-69 covers the 40-69 years range
grouped_data["40-69 years"] += count
else: # 70+ years
grouped_data["70+ years"] += count
# Calculate total
total = sum(grouped_data.values())
# Convert to DataFrame for easier handling
df = pd.DataFrame(list(grouped_data.items()), columns=['Age Group', 'Count'])
df['Percentage'] = (df['Count'] / total * 100).round(1)
df['Label'] = df.apply(lambda x: f"{x['Age Group']}\n{x['Count']:,} ({x['Percentage']}%)", axis=1)
# Create figure
plt.figure(figsize=(10, 6))
# Create color map with shades of teal
teal_shades = ['#006d77', '#83c5be', '#edf6f9'] # Different shades of teal
# Create treemap
squarify.plot(
sizes=df['Count'],
label=df['Label'],
color=teal_shades,
alpha=0.8,
pad=0.02,
text_kwargs={'fontsize': 12}
)
plt.axis('off')
plt.title('Emigration by Life Stage', fontsize=16)
plt.tight_layout()
plt.savefig('emigration_lifestage_treemap.png', dpi=300, bbox_inches='tight')
plt.show()
Looking at the change in emigration between 2023 and 2024, 9% more under 20s stayed but 6% fewer working age adults stayed.
This could be that fewer Ukrainian families returned last year.
In [4]:
#source UDVAN because it has foreign/Denmark as an option so you don't have to aggregate all the citizenships.
destinations_df = pd.read_csv('destinations2024.csv', header=None, encoding='ISO-8859-1')
destinations_df.columns = ['Citizenship', 'Destination Country', 'Emigrant Count']
destinations_df = destinations_df.drop('Citizenship', axis=1)
# Ensure 'Emigrant Count' is numeric, and coerce any non-numeric values to NaN
destinations_df['Emigrant Count'] = pd.to_numeric(destinations_df['Emigrant Count'], errors='coerce')
# Drop rows with a 0 or NaN 'Emigrant Count'
destinations_df = destinations_df[destinations_df['Emigrant Count'].notna() & (destinations_df['Emigrant Count'] != 0)]
# Sort by 'Emigrant Count' in descending order
destinations_df = destinations_df.sort_values(by='Emigrant Count', ascending=False)
# Reset index
destinations_df = destinations_df.reset_index(drop=True)
# Select the top 15 rows
top_15 = destinations_df.head(15)
top_15.head(15)
Out[4]:
| Destination Country | Emigrant Count | |
|---|---|---|
| 0 | Not stated | 11223 |
| 1 | USA | 3800 |
| 2 | Romania | 3621 |
| 3 | Poland | 3024 |
| 4 | Germany | 3001 |
| 5 | Ukraine | 2363 |
| 6 | Sweden | 1767 |
| 7 | Norway | 1638 |
| 8 | Spain | 1510 |
| 9 | Italy | 1376 |
| 10 | India | 1119 |
| 11 | United Kingdom | 1100 |
| 12 | France | 1079 |
| 13 | Lithuania | 1016 |
| 14 | Netherlands | 1012 |
In [5]:
# Create the figure and axis
fig, ax = plt.subplots(figsize=(10, 6))
colors = ['#0092B7']
categories=top_15['Destination Country']
values=top_15['Emigrant Count']
# Create a bar chart
ax.barh(categories, values, color=colors)
# Invert the y-axis to have the largest bar at the top
plt.gca().invert_yaxis()
# Add labels and title
ax.set_xlabel('Destination')
ax.set_title('Popularity of Destination Countries for Foreign Emigrants in 2024')
# Remove x-axis labels
ax.xaxis.set_visible(False)
# Add values on the bars for clarity
for i, v in enumerate(values):
ax.text(v + 1000, i, str(v), color='black', va='center')
# Remove the borders (spines)
for spine in ax.spines.values():
spine.set_visible(False)
# Save the chart as a PNG file
plt.savefig('destinations2024.png', bbox_inches='tight')
# Show the chart
plt.show()
I am curious about the "not stated" category, if it is hiding a very popular destination for example.
2024¶
- USA 3800 ⬆️ (and increased by 7% since last year)
- Romania 3621 ⬆️ (and increased by 23% since last year)
- Poland 3024 ⬆️ (and increased by 23% since last year)
- Germany 3001
- Ukraine 2363 ⬇️ (decreased by 43% since last year)
- Sweden 1767
- Norway 1638
- Spain 1510
- Italy 1376
- India 1119 ⬆️
- United Kingdom 1100 ⬇️
- France 1079 ⬆️
- Lithuania 1016 ⬇️
- Netherlands 1012
In [6]:
# Source: VAN2AAR on https://statbank.dk/
emigration_2024_df = pd.read_csv('emigration 2024.csv', header=None, encoding='ISO-8859-1')
emigration_2024_df.columns = ['Citizenship', 'Destination Country', 'Emigrant Count']
emigration_2024_df = emigration_2024_df[emigration_2024_df['Emigrant Count'] != 0]
emigration_2024_df = emigration_2024_df[emigration_2024_df['Destination Country'].str.lower() != 'not stated']
emigration_2024_df = emigration_2024_df[emigration_2024_df['Citizenship'].str.lower() != 'not stated']
emigration_2024_df = emigration_2024_df[emigration_2024_df['Citizenship'].str.lower() != 'stateless']
In [ ]:
# Keep rows where the 'Citizenship' is the same as the 'Destination Country'
citizen = emigration_2024_df[emigration_2024_df['Citizenship'].str.lower() == emigration_2024_df['Destination Country'].str.lower()]
# Reset the index to avoid gaps after dropping rows
citizen = citizen.reset_index(drop=True)
citizen = citizen.drop('Destination Country', axis=1)
# Sort the ranking
citizen = citizen.sort_values(by='Emigrant Count', ascending=False)
# Reset the index to avoid gaps after dropping rows
citizen = citizen.reset_index(drop=True)
citizen.head(5)
Out[ ]:
| Citizenship | Emigrant Count | |
|---|---|---|
| 0 | Romania | 3558 |
| 1 | USA | 3202 |
| 2 | Poland | 2829 |
| 3 | Ukraine | 2338 |
| 4 | Germany | 2089 |
So the million dollar question is, how many people are "going home"? For those who are going to start again in another country, which countries are popular?
In [8]:
# Drop rows where the 'Citizenship' is the same as the 'Destination Country'
emigration_2024_df = emigration_2024_df[emigration_2024_df['Citizenship'].str.lower() != emigration_2024_df['Destination Country'].str.lower()]
# Reset the index to avoid gaps after dropping rows
emigration_2024_df = emigration_2024_df.reset_index(drop=True)
# Group by destination country and sum the emigrant count
destination_ranking_df = emigration_2024_df.groupby('Destination Country')['Emigrant Count'].sum().reset_index()
# Sort by emigrant count in descending order (highest emigration first)
destination_ranking_df = destination_ranking_df.sort_values(by='Emigrant Count', ascending=False)
destination_ranking_df = destination_ranking_df.reset_index(drop=True)
# If you want to save it to a CSV file
#destination_ranking.to_csv('destination_ranking_2024.csv', index=False)
destination_ranking_df.head(15)
Out[8]:
| Destination Country | Emigrant Count | |
|---|---|---|
| 0 | Germany | 906 |
| 1 | Sweden | 668 |
| 2 | Spain | 605 |
| 3 | USA | 596 |
| 4 | United Kingdom | 498 |
| 5 | Netherlands | 429 |
| 6 | Norway | 331 |
| 7 | Argentina | 296 |
| 8 | France | 288 |
| 9 | Switzerland | 287 |
| 10 | Portugal | 241 |
| 11 | Australia | 240 |
| 12 | Canada | 237 |
| 13 | Italy | 222 |
| 14 | Poland | 195 |
Another question is: which nationalities are serial expats?
In [9]:
# Group by citizenship and sum the emigrant count
citizenship_ranking = emigration_2024_df.groupby('Citizenship')['Emigrant Count'].sum().reset_index()
# Sort by emigrant count in descending order (highest emigration first)
citizenship_ranking = citizenship_ranking.sort_values(by='Emigrant Count', ascending=False)
citizenship_ranking = citizenship_ranking.reset_index(drop=True)
citizenship_ranking.head(10)
Out[9]:
| Citizenship | Emigrant Count | |
|---|---|---|
| 0 | Italy | 828 |
| 1 | Ukraine | 595 |
| 2 | Germany | 452 |
| 3 | Romania | 383 |
| 4 | India | 313 |
| 5 | Spain | 286 |
| 6 | Poland | 247 |
| 7 | China | 243 |
| 8 | Sweden | 241 |
| 9 | France | 241 |
In [10]:
# Path to the GeoJSON file for country boundaries
geojson_url = 'https://raw.githubusercontent.com/python-visualization/folium/master/examples/data/world-countries.json'
# Create a mapping of country names to align them with GeoJSON names
country_name_mapping = {
'USA': 'United States of America',
'UK': 'United Kingdom',
'Russia': 'Russian Federation',
# Add other mappings as needed...
}
# Apply the mapping
destination_ranking_df['Destination Country'] = destination_ranking_df['Destination Country'].replace(country_name_mapping)
# Initialize the Folium map centered on Denmark
m = folium.Map(location=[56.26392, 9.501785], zoom_start=1)
mean = np.mean(destination_ranking_df['Emigrant Count'])
std_dev = np.std(destination_ranking_df['Emigrant Count'])
# Define threshold scale based on standard deviations
threshold_scale = [
0, # Minimum threshold
max(1, mean - std_dev), # At least 1 emigrant, avoid negative or too small values
mean, # Mean threshold
mean + std_dev, # 1 standard deviation above the mean
mean + 2 * std_dev, # 2 standard deviations above the mean
destination_ranking_df['Emigrant Count'].max() # Max value from the data
]
# Ensure that threshold values are strictly increasing
threshold_scale = sorted(list(set(threshold_scale)))
# Add Choropleth layer for countries, using Emigrant Count to define the color
folium.Choropleth(
geo_data=geojson_url,
name='choropleth',
data=destination_ranking_df,
columns=['Destination Country', 'Emigrant Count'],
key_on='feature.properties.name', # Match country names in GeoJSON
fill_color='YlGnBu',
fill_opacity=0.7,
line_opacity=0.2,
legend_name='Emigrant Count',
nan_fill_opacity=0, # Ensure countries not in the data remain unchanged
threshold_scale=threshold_scale,
#threshold_scale=[0, 100, 400, 600, 700, 800],
).add_to(m)
# Add layer control to toggle
folium.LayerControl().add_to(m)
# Save the map
#m.save('colored_countries_map_2024.html')
m
Out[10]:
Make this Notebook Trusted to load map: File -> Trust Notebook
Looking at Region Midtjylland¶
Source: VAN2AAR on https://statbank.dk/¶
Finding out the destinations and citizenships of "serial expats" from Midtjylland
In [11]:
midt_df = pd.read_csv('midt2024.csv', header=None, encoding='ISO-8859-1')
midt_df.columns = ['Region', 'Citizenship', 'Destination Country', 'Emigrant Count']
midt_df = midt_df.drop('Region', axis=1)
midt_df = midt_df[midt_df['Emigrant Count'] != 0]
midt_df = midt_df[midt_df['Destination Country'].str.lower() != 'not stated']
midt_df = midt_df[midt_df['Citizenship'].str.lower() != 'not stated']
midt_df = midt_df[midt_df['Citizenship'].str.lower() != 'stateless']
# Drop rows where the 'Citizenship' is the same as the 'Destination Country'
midt_df = midt_df[midt_df['Citizenship'].str.lower() != midt_df['Destination Country'].str.lower()]
# Reset the index to avoid gaps after dropping rows
midt_df = midt_df.reset_index(drop=True)
midt_df.head()
# Group by destination country and sum the emigrant count
midt_destination_ranking = midt_df.groupby('Destination Country')['Emigrant Count'].sum().reset_index()
In [12]:
# Group by citizenship and sum the emigrant count
midt_citizenship_ranking = midt_df.groupby('Citizenship')['Emigrant Count'].sum().reset_index()
# Sort by emigrant count in descending order (highest emigration first)
midt_citizenship_ranking = midt_citizenship_ranking.sort_values(by='Emigrant Count', ascending=False)
midt_citizenship_ranking = midt_citizenship_ranking.reset_index(drop=True)
midt_citizenship_ranking.head(10)
Out[12]:
| Citizenship | Emigrant Count | |
|---|---|---|
| 0 | Ukraine | 137 |
| 1 | Italy | 108 |
| 2 | Romania | 105 |
| 3 | Syria | 67 |
| 4 | Germany | 50 |
| 5 | Sweden | 47 |
| 6 | Lithuania | 46 |
| 7 | China | 44 |
| 8 | Spain | 41 |
| 9 | Poland | 41 |
In [13]:
# Group by destination country and sum the emigrant count
midt_destination_ranking = midt_df.groupby('Destination Country')['Emigrant Count'].sum().reset_index()
# Sort by emigrant count in descending order (highest emigration first)
midt_destination_ranking = midt_destination_ranking.sort_values(by='Emigrant Count', ascending=False)
midt_destination_ranking = midt_destination_ranking.reset_index(drop=True)
midt_destination_ranking.head(15)
Out[13]:
| Destination Country | Emigrant Count | |
|---|---|---|
| 0 | Germany | 216 |
| 1 | United Kingdom | 106 |
| 2 | Netherlands | 86 |
| 3 | Spain | 83 |
| 4 | USA | 59 |
| 5 | Poland | 48 |
| 6 | Norway | 46 |
| 7 | Moldova | 44 |
| 8 | Sweden | 43 |
| 9 | Argentina | 40 |
| 10 | Italy | 36 |
| 11 | France | 34 |
| 12 | Switzerland | 31 |
| 13 | Australia | 31 |
| 14 | Lebanon | 30 |
In [14]:
# Apply the mapping to the data
midt_destination_ranking['Destination Country'] = midt_destination_ranking['Destination Country'].replace(country_name_mapping)
# Initialize the Folium map centered on Denmark
m = folium.Map(location=[56.26392, 9.501785], zoom_start=1)
mean = np.mean(midt_destination_ranking['Emigrant Count'])
std_dev = np.std(midt_destination_ranking['Emigrant Count'])
# Define threshold scale based on standard deviations
threshold_scale = [
0, # Minimum threshold
max(1, mean - std_dev), # At least 1 emigrant, avoid negative or too small values
mean, # Mean threshold
mean + std_dev, # 1 standard deviation above the mean
mean + 2 * std_dev, # 2 standard deviations above the mean
midt_destination_ranking['Emigrant Count'].max() # Max value from the data
]
# Ensure that threshold values are strictly increasing
threshold_scale = sorted(list(set(threshold_scale)))
# Add Choropleth layer for countries, using Emigrant Count to define the color
folium.Choropleth(
geo_data=geojson_url,
name='choropleth',
data=midt_destination_ranking,
columns=['Destination Country', 'Emigrant Count'],
key_on='feature.properties.name', # Match country names in GeoJSON
fill_color='YlGnBu',
fill_opacity=0.7,
line_opacity=0.2,
legend_name='Emigrant Count',
nan_fill_opacity=0, # Ensure countries not in the data remain unchanged
threshold_scale=threshold_scale,
#threshold_scale=[0, 100, 400, 600, 700, 800],
).add_to(m)
# Add layer control to toggle
folium.LayerControl().add_to(m)
# Save the map
#m.save('midt_colored_countries_map_2024.html')
m
Out[14]:
Make this Notebook Trusted to load map: File -> Trust Notebook
Looking at Aarhus¶
Source: VAN2AAR on https://statbank.dk/¶
Finding out the destinations and citizenships of emigrating internationals from Aarhus
In [15]:
aarhus_df = pd.read_csv('aarhus2024.csv', header=None, encoding='ISO-8859-1')
aarhus_df.columns = ['City', 'Citizenship', 'Destination Country', 'Emigrant Count']
aarhus_df = aarhus_df.drop('City', axis=1)
aarhus_df = aarhus_df[aarhus_df['Emigrant Count'] != 0]
aarhus_df = aarhus_df[aarhus_df['Destination Country'].str.lower() != 'not stated']
aarhus_df = aarhus_df[aarhus_df['Citizenship'].str.lower() != 'not stated']
aarhus_df = aarhus_df[aarhus_df['Citizenship'].str.lower() != 'stateless']
In [16]:
aarhus_citizen = aarhus_df.drop('Destination Country', axis=1)
# Sort the ranking
aarhus_citizen = aarhus_citizen.sort_values(by='Emigrant Count', ascending=False)
# Reset the index to avoid gaps after dropping rows
aarhus_citizen = aarhus_citizen.reset_index(drop=True)
aarhus_citizen.head(5)
Out[16]:
| Citizenship | Emigrant Count | |
|---|---|---|
| 0 | Germany | 231 |
| 1 | Norway | 209 |
| 2 | China | 102 |
| 3 | Poland | 100 |
| 4 | Romania | 97 |
In [17]:
#Make a dataframe where people go back to their home country
aarhus_home_df = aarhus_df[aarhus_df['Citizenship'].str.lower() == aarhus_df['Destination Country'].str.lower()]
# Sort the ranking
aarhus_home_df = aarhus_home_df.sort_values(by='Emigrant Count', ascending=False)
# Reset the index to avoid gaps after dropping rows
aarhus_home_df = aarhus_home_df.reset_index(drop=True)
# Drop rows where the 'Citizenship' is the same as the 'Destination Country'
aarhus_df = aarhus_df[aarhus_df['Citizenship'].str.lower() != aarhus_df['Destination Country'].str.lower()]
# Reset the index to avoid gaps after dropping rows
aarhus_df = aarhus_df.reset_index(drop=True)
# Group by destination country and sum the emigrant count
aarhus_destination_ranking = aarhus_df.groupby('Destination Country')['Emigrant Count'].sum().reset_index()
In [18]:
aarhus_home_df.head(20)
Out[18]:
| Citizenship | Destination Country | Emigrant Count | |
|---|---|---|---|
| 0 | Germany | Germany | 231 |
| 1 | Norway | Norway | 209 |
| 2 | China | China | 102 |
| 3 | Poland | Poland | 100 |
| 4 | Romania | Romania | 97 |
| 5 | Spain | Spain | 92 |
| 6 | Ukraine | Ukraine | 89 |
| 7 | Italy | Italy | 87 |
| 8 | France | France | 81 |
| 9 | Netherlands | Netherlands | 80 |
| 10 | Iceland | Iceland | 74 |
| 11 | USA | USA | 71 |
| 12 | India | India | 56 |
| 13 | Sweden | Sweden | 51 |
| 14 | Belgium | Belgium | 50 |
| 15 | Japan | Japan | 50 |
| 16 | United Kingdom | United Kingdom | 48 |
| 17 | Hungary | Hungary | 45 |
| 18 | Australia | Australia | 40 |
| 19 | Canada | Canada | 30 |
In [19]:
# Sort by emigrant count in descending order (highest emigration first)
aarhus_destination_ranking = aarhus_destination_ranking.sort_values(by='Emigrant Count', ascending=False)
aarhus_destination_ranking = aarhus_destination_ranking.reset_index(drop=True)
# If you want to save it to a CSV file
#aarhus_destination_ranking.to_csv('aarhus_destination_ranking.csv', index=False)
aarhus_destination_ranking.head(15)
Out[19]:
| Destination Country | Emigrant Count | |
|---|---|---|
| 0 | Germany | 104 |
| 1 | United Kingdom | 72 |
| 2 | Netherlands | 63 |
| 3 | Spain | 60 |
| 4 | Argentina | 26 |
| 5 | Switzerland | 25 |
| 6 | Sweden | 25 |
| 7 | France | 24 |
| 8 | Norway | 22 |
| 9 | Italy | 22 |
| 10 | Australia | 21 |
| 11 | USA | 20 |
| 12 | Belgium | 14 |
| 13 | Czech Republic | 12 |
| 14 | Ireland | 12 |
In [20]:
# Group by citizenship and sum the emigrant count
aarhus_citizenship_ranking = aarhus_df.groupby('Citizenship')['Emigrant Count'].sum().reset_index()
# Sort by emigrant count in descending order (highest emigration first)
aarhus_citizenship_ranking = aarhus_citizenship_ranking.sort_values(by='Emigrant Count', ascending=False)
aarhus_citizenship_ranking = aarhus_citizenship_ranking.reset_index(drop=True)
aarhus_citizenship_ranking.head(10)
Out[20]:
| Citizenship | Emigrant Count | |
|---|---|---|
| 0 | Italy | 93 |
| 1 | China | 35 |
| 2 | Germany | 29 |
| 3 | Spain | 28 |
| 4 | Ukraine | 27 |
| 5 | Romania | 26 |
| 6 | India | 24 |
| 7 | Sweden | 22 |
| 8 | Poland | 19 |
| 9 | Hungary | 17 |
In [21]:
# Apply the name mapping
aarhus_destination_ranking['Destination Country'] = aarhus_destination_ranking['Destination Country'].replace(country_name_mapping)
# Initialize the Folium map centered on Denmark
m = folium.Map(location=[56.26392, 9.501785], zoom_start=1)
mean = np.mean(midt_destination_ranking['Emigrant Count'])
std_dev = np.std(midt_destination_ranking['Emigrant Count'])
# Define threshold scale based on standard deviations
threshold_scale = [
0, # Minimum threshold
max(1, mean - std_dev), # At least 1 emigrant, avoid negative or too small values
mean, # Mean threshold
mean + std_dev, # 1 standard deviation above the mean
mean + 2 * std_dev, # 2 standard deviations above the mean
aarhus_destination_ranking['Emigrant Count'].max() # Max value from the data
]
# Ensure that threshold values are strictly increasing
threshold_scale = sorted(list(set(threshold_scale)))
# Add Choropleth layer for countries, using Emigrant Count to define the color
folium.Choropleth(
geo_data=geojson_url,
name='choropleth',
data=aarhus_destination_ranking,
columns=['Destination Country', 'Emigrant Count'],
key_on='feature.properties.name', # Match country names in GeoJSON
fill_color='YlGnBu',
fill_opacity=0.7,
line_opacity=0.2,
legend_name='Emigrant Count',
nan_fill_opacity=0, # Ensure countries not in the data remain unchanged
threshold_scale=threshold_scale,
#threshold_scale=[0, 100, 400, 600, 700, 800],
).add_to(m)
# Add layer control to toggle
folium.LayerControl().add_to(m)
# Save the map
m.save('aarhus_colored_countries_map_2024.html')
m
Out[21]:
Make this Notebook Trusted to load map: File -> Trust Notebook
In [22]:
# Load the data
df_aarhus_age = pd.read_csv('aarhus_age_24.csv', header=None, encoding='ISO-8859-1')
df_aarhus_age.columns = ['City', 'Age', 'Citizenship', 'Emigrant Count']
df_aarhus_age = df_aarhus_age.drop('City', axis=1)
# Clean the Age column by removing "years"/"year" and converting to integer
def clean_age(age_str):
# Convert to string to ensure string methods work
age_str = str(age_str)
# Remove "years" or "year" and any extra spaces
age_str = age_str.replace('years', '').replace('year', '').strip()
# Handle any "+" symbol for ranges like "110+"
if '+' in age_str:
age_str = age_str.replace('+', '').strip()
# Convert to integer
try:
return int(age_str)
except ValueError:
# If conversion fails, return original value
print(f"Could not convert {age_str} to integer")
return age_str
# Apply the cleaning function to the Age column
df_aarhus_age['Age'] = df_aarhus_age['Age'].apply(clean_age)
# Now the Age column should be an integer and will sort correctly
df_aarhus_age_summed = df_aarhus_age.groupby('Age')['Emigrant Count'].sum().reset_index()
# Drop ages with 0 emigrants
df_aarhus_age_summed = df_aarhus_age_summed[df_aarhus_age_summed['Emigrant Count'] > 0]
print(df_aarhus_age_summed)
Age Emigrant Count 0 0 16 1 1 24 2 2 23 3 3 22 4 4 18 .. ... ... 78 78 1 79 79 1 80 80 1 83 83 1 92 92 1 [82 rows x 2 columns]
In [23]:
# Group into age ranges
grouped_data = {"0-20 years": 0, "21-39 years": 0, "40-69 years": 0, "70+ years": 0}
# Iterate through rows of the DataFrame
for index, row in df_aarhus_age_summed.iterrows():
age = row['Age']
count = row['Emigrant Count']
# Convert age to integer if it's not already
try:
age_num = int(age)
except ValueError:
# Handle cases like "110 years +" if they exist
if "+" in str(age):
age_num = int(str(age).split()[0])
else:
continue # Skip if we can't parse the age
# Categorize based on age groups
if age_num <= 20:
grouped_data["0-20 years"] += count
elif age_num <= 39:
grouped_data["21-39 years"] += count
elif age_num <= 69:
grouped_data["40-69 years"] += count
else: # 70+ years
grouped_data["70+ years"] += count
In [24]:
# Calculate total, ensure it's not zero
total = sum(grouped_data.values())
if total == 0:
print("Warning: Total count is zero!")
total = 1 # Prevent division by zero
# Convert to DataFrame for easier handling
df = pd.DataFrame(list(grouped_data.items()), columns=['Age Group', 'Count'])
# Filter out zero counts before plotting
df_non_zero = df[df['Count'] > 0].copy()
if df_non_zero.empty:
print("No data to plot! All counts are zero.")
else:
# Calculate percentages only for non-zero rows
df_non_zero['Percentage'] = (df_non_zero['Count'] / total * 100).round(1)
df_non_zero['Label'] = df_non_zero.apply(lambda x: f"{x['Age Group']}\n{x['Count']:,} ({x['Percentage']}%)", axis=1)
# Create figure
plt.figure(figsize=(10, 6))
# Create color map with shades of teal - adjust colors to match number of groups
teal_shades = ['#006d77', '#83c5be', '#edf6f9', '#57a8a3'][:len(df_non_zero)]
# Create treemap
squarify.plot(
sizes=df_non_zero['Count'],
label=df_non_zero['Label'],
color=teal_shades,
alpha=0.8,
pad=0.02,
text_kwargs={'fontsize': 12}
)
plt.axis('off')
plt.title('Emigration from Aarhus in 2024 by Life Stage', fontsize=16)
plt.tight_layout()
plt.savefig('emigration_lifestage_treemap_aarhus.png', dpi=300, bbox_inches='tight')
plt.show()
In [25]:
# Data from UDVAN on statbank.dk
df_age = pd.read_csv('aarhus_age_buckets_24.csv', header=None, encoding='ISO-8859-1')
df_age.columns = ['Citizenship', 'Age Bracket', 'Emigrant Count']
df_age = df_age.drop('Citizenship', axis=1)
# Ensure 'Emigrant Count' is numeric, and coerce any non-numeric values to NaN
df_age['Emigrant Count'] = pd.to_numeric(df_age['Emigrant Count'], errors='coerce')
# Drop rows with a 0 or NaN 'Emigrant Count'
df_age = df_age[df_age['Emigrant Count'].notna() & (df_age['Emigrant Count'] != 0)]
categories = df_age['Age Bracket']
values = df_age['Emigrant Count']
# Create the figure and axis
fig, ax = plt.subplots(figsize=(10, 6))
colors = ['#0093B7']
# Create a horizontal bar chart
ax.barh(categories, values, color=colors)
# Add labels and title
ax.set_xlabel('Number of People')
ax.set_ylabel('Age Groups')
ax.set_title('Emigrants from Aarhus by Age Group in 2024')
ax.xaxis.set_visible(True)
ax.tick_params("x", rotation=30)
# Add values on the bars for clarity
#for i, v in enumerate(values):
# ax.text(v + 1000, i, str(v), color='black', va='center')
# Remove the borders (spines)
#for spine in ax.spines.values():
# spine.set_visible(False)
# Save the chart as a PNG file
plt.savefig('aarhus_age.png', bbox_inches='tight')
# Show the chart
plt.tight_layout()
plt.show()
