Natural Disasters Analysis - Yearly Trend

Authors: Women of the West Coast (WWC)

Date: Oct. 22, 2023

Table of Contents

1. Introduction
2. Data Processing
3. Global Analysis
4. Canada-Specific Analysis
5. Insights Summary
6. Conclusion

Introduction

Natural disasters have been causing significant socio-economic impacts around the globe. Understanding the patterns, economic damages, and loss of life due to these events is crucial for effective crisis response and management. In this report, we analyze global and Canada-specific data on natural disasters to gain insights that can inform and empower response initiatives.

Data Processing

The datasets provided contained information about various natural disasters across different years. We combined these datasets, ensuring consistency in column names and data types. Duplicate and missing values were addressed to ensure the integrity and reliability of the subsequent analysis.

import pandas as pd
import matplotlib.pyplot as plt

# Filter warnings
from warnings import filterwarnings
filterwarnings('ignore')

# Load the two datasets
dataset_1 = pd.read_csv('1900_2021_DISASTERS.xlsx - emdat data.csv')
dataset_2 = pd.read_csv('1970-2021_DISASTERS.xlsx - emdat data.csv')

# Set display options to show all columns and rows
pd.set_option('display.max_columns', None)
pd.set_option('display.max_rows', None)

# Display the first few rows of dataset_1
dataset_1.head()

	Year	Seq	Glide	Disaster Group	Disaster Subgroup	Disaster Type	Disaster Subtype	Disaster Subsubtype	Event Name	Country	ISO	Region	Continent	Location	Origin	Associated Dis	Associated Dis2	OFDA Response	Appeal	Declaration	Aid Contribution	Dis Mag Value	Dis Mag Scale	Latitude	Longitude	Local Time	River Basin	Start Year	Start Month	Start Day	End Year	End Month	End Day	Total Deaths	No Injured	No Affected	No Homeless	Total Affected	Insured Damages ('000 US$)	Total Damages ('000 US$)	CPI	Adm Level	Admin1 Code	Admin2 Code	Geo Locations
0	1900	9002	NaN	Natural	Climatological	Drought	Drought	NaN	NaN	Cabo Verde	CPV	Western Africa	Africa	Countrywide	NaN	Famine	NaN	NaN	No	No	NaN	NaN	Km2	NaN	NaN	NaN	NaN	1900	NaN	NaN	1900	NaN	NaN	11000.0	NaN	NaN	NaN	NaN	NaN	NaN	3.221647	NaN	NaN	NaN	NaN
1	1900	9001	NaN	Natural	Climatological	Drought	Drought	NaN	NaN	India	IND	Southern Asia	Asia	Bengal	NaN	NaN	NaN	NaN	No	No	NaN	NaN	Km2	NaN	NaN	NaN	NaN	1900	NaN	NaN	1900	NaN	NaN	1250000.0	NaN	NaN	NaN	NaN	NaN	NaN	3.221647	NaN	NaN	NaN	NaN
2	1902	12	NaN	Natural	Geophysical	Earthquake	Ground movement	NaN	NaN	Guatemala	GTM	Central America	Americas	Quezaltenango, San Marcos	NaN	Tsunami/Tidal wave	NaN	NaN	NaN	NaN	NaN	8.0	Richter	14	-91	20:20	NaN	1902	4.0	18.0	1902	4.0	18.0	2000.0	NaN	NaN	NaN	NaN	NaN	25000.0	3.350513	NaN	NaN	NaN	NaN
3	1902	3	NaN	Natural	Geophysical	Volcanic activity	Ash fall	NaN	Santa Maria	Guatemala	GTM	Central America	Americas	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1902	4.0	8.0	1902	4.0	8.0	1000.0	NaN	NaN	NaN	NaN	NaN	NaN	3.350513	NaN	NaN	NaN	NaN
4	1902	10	NaN	Natural	Geophysical	Volcanic activity	Ash fall	NaN	Santa Maria	Guatemala	GTM	Central America	Americas	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1902	10.0	24.0	1902	10.0	24.0	6000.0	NaN	NaN	NaN	NaN	NaN	NaN	3.350513	NaN	NaN	NaN	NaN

Dataset 1:

Covers the years from 1900 onwards. Contains 45 columns, including details like the year, disaster type, country, number of deaths, economic damages, and more.

# Display the first few rows of dataset_2
dataset_2.head()

	Dis No	Year	Seq	Glide	Disaster Group	Disaster Subgroup	Disaster Type	Disaster Subtype	Disaster Subsubtype	Event Name	Country	ISO	Region	Continent	Location	Origin	Associated Dis	Associated Dis2	OFDA Response	Appeal	Declaration	Aid Contribution	Dis Mag Value	Dis Mag Scale	Latitude	Longitude	Local Time	River Basin	Start Year	Start Month	Start Day	End Year	End Month	End Day	Total Deaths	No Injured	No Affected	No Homeless	Total Affected	Reconstruction Costs ('000 US$)	Insured Damages ('000 US$)	Total Damages ('000 US$)	CPI	Adm Level	Admin1 Code	Admin2 Code	Geo Locations
0	1970-0013-ARG	1970	13	NaN	Natural	Hydrological	Flood	NaN	NaN	NaN	Argentina	ARG	South America	Americas	Mendoza	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Km2	NaN	NaN	NaN	NaN	1970	1.0	4.0	1970	1.0	4.0	36.0	NaN	NaN	NaN	NaN	NaN	NaN	25000.0	15.001282	NaN	NaN	NaN	NaN
1	1970-0109-AUS	1970	109	NaN	Natural	Meteorological	Storm	Tropical cyclone	NaN	Ada	Australia	AUS	Australia and New Zealand	Oceania	Queensland	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Kph	NaN	NaN	NaN	NaN	1970	1.0	NaN	1970	1.0	NaN	13.0	NaN	NaN	NaN	NaN	NaN	NaN	72475.0	15.001282	NaN	NaN	NaN	NaN
2	1970-0044-BEN	1970	44	NaN	Natural	Hydrological	Flood	NaN	NaN	NaN	Benin	BEN	Western Africa	Africa	Atacora region	NaN	NaN	NaN	Yes	NaN	NaN	NaN	NaN	Km2	NaN	NaN	NaN	NaN	1970	9.0	NaN	1970	9.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	200.0	15.001282	NaN	NaN	NaN	NaN
3	1970-0063-BGD	1970	63	NaN	Natural	Meteorological	Storm	Tropical cyclone	NaN	NaN	Bangladesh	BGD	Southern Asia	Asia	Khulna, Chittagong	NaN	NaN	NaN	Yes	NaN	NaN	NaN	NaN	Kph	NaN	NaN	NaN	NaN	1970	11.0	12.0	1970	11.0	12.0	300000.0	NaN	3648000.0	NaN	3648000.0	NaN	NaN	86400.0	15.001282	NaN	NaN	NaN	NaN
4	1970-0026-BGD	1970	26	NaN	Natural	Meteorological	Storm	NaN	NaN	NaN	Bangladesh	BGD	Southern Asia	Asia	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Kph	NaN	NaN	NaN	NaN	1970	4.0	13.0	1970	4.0	13.0	17.0	NaN	110.0	NaN	110.0	NaN	NaN	NaN	15.001282	NaN	NaN	NaN	NaN

Dataset 2:

Covers the years from 1970 onwards. Contains 47 columns, with most columns overlapping with Dataset 1 and a few additional ones.

Combine datasets

To effectively combine and analyze the datasets, we'll need to:

1. Identify and handle overlapping years between the datasets.
2. Ensure that the column names and data types are consistent.
3. Merge the datasets based on common columns.

Let's start by examining the common and unique columns between the two datasets.

# Identify common and unique columns between the two datasets
common_columns = set(dataset_1.columns).intersection(set(dataset_2.columns))
unique_to_dataset_1 = set(dataset_1.columns) - set(dataset_2.columns)
unique_to_dataset_2 = set(dataset_2.columns) - set(dataset_1.columns)

common_columns, unique_to_dataset_1, unique_to_dataset_2

({'Adm Level',
  'Admin1 Code',
  'Admin2 Code',
  'Aid Contribution',
  'Appeal',
  'Associated Dis',
  'Associated Dis2',
  'CPI',
  'Continent',
  'Country',
  'Declaration',
  'Dis Mag Scale',
  'Dis Mag Value',
  'Disaster Group',
  'Disaster Subgroup',
  'Disaster Subsubtype',
  'Disaster Subtype',
  'Disaster Type',
  'End Day',
  'End Month',
  'End Year',
  'Event Name',
  'Geo Locations',
  'Glide',
  'ISO',
  "Insured Damages ('000 US$)",
  'Latitude',
  'Local Time',
  'Location',
  'Longitude',
  'No Affected',
  'No Homeless',
  'No Injured',
  'OFDA Response',
  'Origin',
  'Region',
  'River Basin',
  'Seq',
  'Start Day',
  'Start Month',
  'Start Year',
  'Total Affected',
  "Total Damages ('000 US$)",
  'Total Deaths',
  'Year'},
 set(),
 {'Dis No', "Reconstruction Costs ('000 US$)"})

Here's a summary of the columns in the two datasets:

Common Columns (between both datasets):

Both datasets share 45 columns in common, such as 'Year', 'Disaster Type', 'Country', 'Total Deaths', 'Total Damages ('000 US$)', and more.

Unique Columns:

• Dataset 1 doesn't have any unique columns. Every column in Dataset 1 is also present in Dataset 2.
• Dataset 2 has two unique columns: 'Dis No' and 'Reconstruction Costs ('000 US$)'.

Given this information, our approach will be to:

1. Combine the datasets based on the common columns.
2. Retain the unique columns from Dataset 2 in the combined dataset.
3. Handle overlapping years between the datasets to avoid duplicate entries.

Let's proceed with the combination.

# Combining the datasets based on common columns
combined_dataset = pd.concat([dataset_1[list(common_columns)], dataset_2[list(common_columns)]], axis=0)

# Sorting the combined dataset based on Year and then by Seq (sequence number) for better organization
combined_dataset = combined_dataset.sort_values(by=['Year', 'Seq'])

# Resetting the index for the combined dataset
combined_dataset.reset_index(drop=True, inplace=True)

# Display the first few rows of the combined dataset
combined_dataset_head = combined_dataset.head()

combined_dataset_head

	End Year	Location	Admin1 Code	Disaster Group	ISO	Event Name	Country	Aid Contribution	End Day	OFDA Response	Dis Mag Value	Start Year	Start Day	Total Damages ('000 US$)	Disaster Subgroup	Admin2 Code	Region	Start Month	Latitude	Continent	Glide	Appeal	Associated Dis	No Affected	Seq	Insured Damages ('000 US$)	Disaster Subsubtype	Geo Locations	Dis Mag Scale	Adm Level	Year	Disaster Subtype	CPI	Total Deaths	Total Affected	Longitude	Disaster Type	Declaration	No Injured	Associated Dis2	Local Time	River Basin	Origin	End Month	No Homeless
0	1900	Galveston (Texas)	NaN	Natural	USA	NaN	United States of America (the)	NaN	8.0	NaN	220.0	1900	8.0	30000.0	Meteorological	NaN	Northern America	9.0	NaN	Americas	NaN	NaN	Avalanche (Snow, Debris)	NaN	3	NaN	NaN	NaN	Kph	NaN	1900	Tropical cyclone	3.221647	6000.0	NaN	NaN	Storm	NaN	NaN	NaN	NaN	NaN	NaN	9.0	NaN
1	1900	Saint James	NaN	Natural	JAM	NaN	Jamaica	NaN	6.0	NaN	NaN	1900	6.0	NaN	Hydrological	NaN	Caribbean	1.0	NaN	Americas	NaN	NaN	NaN	NaN	6	NaN	NaN	NaN	Km2	NaN	1900	NaN	3.221647	300.0	NaN	NaN	Flood	NaN	NaN	NaN	NaN	NaN	NaN	1.0	NaN
2	1900	Porus	NaN	Natural	JAM	Gastroenteritis	Jamaica	NaN	13.0	NaN	NaN	1900	13.0	NaN	Biological	NaN	Caribbean	1.0	33	Americas	NaN	NaN	NaN	NaN	7	NaN	NaN	NaN	Vaccinated	NaN	1900	Viral disease	3.221647	30.0	NaN	NaN	Epidemic	NaN	NaN	NaN	NaN	NaN	NaN	1.0	NaN
3	1900	NaN	NaN	Natural	JPN	NaN	Japan	NaN	7.0	NaN	NaN	1900	7.0	NaN	Geophysical	NaN	Eastern Asia	7.0	NaN	Asia	NaN	NaN	NaN	NaN	8	NaN	NaN	NaN	NaN	NaN	1900	Ash fall	3.221647	30.0	NaN	NaN	Volcanic activity	NaN	NaN	NaN	NaN	NaN	NaN	7.0	NaN
4	1900	KARS,KARAKURT,KAGIZMAN,DIGOR	NaN	Natural	TUR	NaN	Turkey	NaN	12.0	NaN	6.0	1900	12.0	NaN	Geophysical	NaN	Western Asia	7.0	40.3	Asia	NaN	NaN	NaN	NaN	9	NaN	NaN	NaN	Richter	NaN	1900	Ground movement	3.221647	140.0	NaN	43.1	Earthquake	NaN	NaN	NaN	09:25	NaN	NaN	7.0	NaN

The datasets have been successfully combined based on the common columns.

Data Cleaning

Before diving into analysis, it's essential to clean the dataset to ensure accurate results.

Checking for Missing Values

We will start by identifying the columns with missing values and counting the number of missing values for each column.

# Identifying columns with missing values and counting the number of missing values for each column
missing_values = combined_dataset.isnull().sum()
missing_values = missing_values[missing_values > 0].sort_values(ascending=False)

missing_values

Aid Contribution              29416
Associated Dis2               29365
Local Time                    28902
Disaster Subsubtype           28649
Insured Damages ('000 US$)    28580
River Basin                   28198
OFDA Response                 27626
Glide                         27608
No Homeless                   26091
Appeal                        25761
Latitude                      25710
Longitude                     25703
Declaration                   24387
Associated Dis                24190
Event Name                    23264
No Injured                    23224
Origin                        23196
Admin2 Code                   22836
Admin1 Code                   21608
Dis Mag Value                 21255
Total Damages ('000 US$)      20662
Adm Level                     15056
Geo Locations                 15056
No Affected                   12704
Total Deaths                   9158
Total Affected                 8112
Start Day                      6695
End Day                        6550
Disaster Subtype               5857
Location                       3138
Dis Mag Scale                  2263
End Month                      1257
Start Month                     655
CPI                             630
dtype: int64

Decision for Dealing with Missing Values

Since natural disasters and crisis situations are often characterized by rapidly evolving scenarios where collecting comprehensive and precise data can be challenging. The presence of missing values in this dataset is a testament to the complexities involved in data collection under such circumstances.

Given the nature of this dataset, imputing or modifying the missing values might not be appropriate. Imputation, in this context, could introduce biases or inaccuracies that are not reflective of the ground realities. Thus, for the sake of maintaining the integrity and authenticity of the insights derived from this dataset, it's prudent to retain the missing values as they are.

By doing so, we acknowledge the gaps in the data and ensure that any subsequent analyses or interpretations consider these gaps. This approach upholds the principle of accurately reflecting the insights from these crisis situations without making assumptions that could potentially distort the narrative.

Next, let's conduct some exploratory data analysis (EDA) to gather insights from the combined dataset.

Global Analysis

We began by understanding the global impact of natural disasters over the years. This included:

• Yearly trends of the number of disasters and total damages.
• Distribution of different disaster types.

# Yearly trend of number of disasters
yearly_disaster_counts = combined_dataset.groupby('Year').size()

# Yearly trend of total damages
yearly_total_damages = combined_dataset.groupby('Year')["Total Damages ('000 US$)"].sum()

plt.figure(figsize=(14, 7))

# Plotting number of disasters
plt.subplot(1, 2, 1)
yearly_disaster_counts.plot(color='blue')
plt.title('Yearly Trend of Number of Disasters', fontweight='bold')
plt.xlabel('Year', fontweight='bold')
plt.ylabel('Number of Disasters', fontweight='bold')
plt.grid(True, which='both', linestyle='--', linewidth=0.5)

# Plotting total damages
plt.subplot(1, 2, 2)
yearly_total_damages.plot(color='red')
plt.title('Yearly Trend of Total Damages from Disasters', fontweight='bold')
plt.xlabel('Year', fontweight='bold')
plt.ylabel('Total Damages (in \'000 US$)', fontweight='bold')
plt.grid(True, which='both', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

The graphs provide the following insights:

1. Yearly Trend of Number of Disasters:

• There's a noticeable increase in the number of reported disasters from the mid-20th century onwards.
• The peak in the number of disasters seems to occur around the early 21st century.

1. Yearly Trend of Total Damages from Disasters:

• The economic damages from natural disasters show significant variability over the years.
• There are several pronounced spikes, especially in the late 20th and early 21st centuries, indicating years of particularly high economic damages.

Next, let's examine the distribution of different disaster types in the dataset to understand the most common types of disasters reported.

# Distribution of disaster types
disaster_type_counts = combined_dataset['Disaster Type'].value_counts()

plt.figure(figsize=(14, 7))
disaster_type_counts.plot(kind='bar', color='skyblue', edgecolor='black')
plt.title('Distribution of Disaster Types', fontweight='bold')
plt.xlabel('Disaster Type', fontweight='bold')
plt.ylabel('Number of Occurrences', fontweight='bold')
plt.xticks(rotation=45)
plt.grid(axis='y', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

The bar chart provides insights into the distribution of disaster types:

• Floods are the most frequently reported disaster type, followed closely by storms.
• Droughts, earthquakes, and extreme temperature events are also common.
• Volcanic activity, wildfires, and landslides are less frequent but still significant.

Understanding the distribution of disaster types can help tailor the platform's features and resources to cater to the most common disaster scenarios.

Canada-Specific Analysis

Now, let's shift our focus to Canada-specific insights. This includes:

• Examine the distribution of different disaster types in Canada.
• Analyze the trend of economic damages in Canada due to natural disasters over the years.

We'll begin by examining the distribution of different disaster types in Canada.

# Extracting Canada-specific data
canada_data = combined_dataset[combined_dataset['Country'] == 'Canada']

# Distribution of disaster types in Canada
canada_disaster_type_counts = canada_data['Disaster Type'].value_counts()

plt.figure(figsize=(14, 7))
canada_disaster_type_counts.plot(kind='bar', color='cadetblue', edgecolor='black')
plt.title('Distribution of Disaster Types in Canada', fontweight='bold')
plt.xlabel('Disaster Type', fontweight='bold')
plt.ylabel('Number of Occurrences', fontweight='bold')
plt.xticks(rotation=45)
plt.grid(axis='y', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

The bar chart displays the distribution of disaster types in Canada:

• Storms and floods are the two most frequently reported disaster types in Canada.
• Wildfires and extreme temperature events also occur frequently.
• Landslides, droughts, and earthquakes are less common but still significant.

Next, we'll delve into the top 4 disaster types in Canada. We'll begin by analyzing the yearly occurrences for each of these disaster types.

# Top 4 disaster types in Canada
top_4_disasters_canada = canada_disaster_type_counts.head(4).index.tolist()

# Creating subplots for yearly occurrences of the top 4 disaster types
plt.figure(figsize=(18, 12))

for i, disaster_type in enumerate(top_4_disasters_canada, 1):
    yearly_occurrences = canada_data[canada_data['Disaster Type'] == disaster_type].groupby('Year').size()
    
    plt.subplot(2, 2, i)
    yearly_occurrences.plot(color='dodgerblue')
    plt.title(f'Yearly Occurrences of {disaster_type} in Canada', fontweight='bold', fontsize=14)
    plt.xlabel('Year', fontsize=12)
    plt.ylabel('Number of Occurrences', fontsize=12)
    plt.grid(True, which='both', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

Graph Interpretations:

1. Yearly Occurrences of Floods in Canada:

• Floods have been consistently occurring in Canada throughout the timeframe from 1900 to 2021.
• The frequency of floods seems to have increased starting from the late 20th century, with several years experiencing multiple flood events.

1. Yearly Occurrences of Storms in Canada:

• Storms also have been a regular occurrence throughout the observed timeframe.
• There's a noticeable increase in the number of storms from the mid-20th century onwards.

1. Yearly Occurrences of Wildfires in Canada:

• Wildfires have been sporadic in the early 20th century, with a clear increase in occurrences from the mid-20th century onwards.
• The early 21st century shows several years with a high frequency of wildfires.

1. Yearly Occurrences of Extreme Temperature in Canada:

• Extreme temperature events were less frequent in the early 20th century but have become more common from the late 20th century onwards.

Next, we'll analyze the trend of economic damages in Canada due to natural disasters over the years. This will provide insights into the economic impact of these events on the country.

# Yearly trend of economic damages in Canada due to natural disasters
yearly_damages_canada = canada_data.groupby('Year')["Total Damages ('000 US$)"].sum()

plt.figure(figsize=(14, 7))
yearly_damages_canada.plot(color='forestgreen')
plt.title('Yearly Trend of Economic Damages from Natural Disasters in Canada', fontweight='bold')
plt.xlabel('Year', fontweight='bold')
plt.ylabel('Economic Damage (in \'000 US$)', fontweight='bold')
plt.grid(True, which='both', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

The graph showcases the yearly trend of economic damages in Canada due to natural disasters:

• From the mid-20th century, there's a clear upward trend in economic damages. The late 20th century and early 21st century have several pronounced spikes, indicating years with particularly high economic impacts.
• These spikes emphasize the increasing economic challenges that natural disasters pose to Canada, necessitating proactive response and management.

Now, let's analyze the economic damages over the years for each of these top 4 disaster types in Canada.

# Creating subplots for yearly economic damages of the top 4 disaster types in Canada
plt.figure(figsize=(18, 12))

for i, disaster_type in enumerate(top_4_disasters_canada, 1):
    yearly_damages_type = canada_data[canada_data['Disaster Type'] == disaster_type].groupby('Year')["Total Damages ('000 US$)"].sum()
    
    plt.subplot(2, 2, i)
    yearly_damages_type.plot(color='darkorchid')
    plt.title(f'Yearly Economic Damages from {disaster_type} in Canada', fontweight='bold', fontsize=14)
    plt.xlabel('Year', fontweight='bold', fontsize=12)
    plt.ylabel('Economic Damage (in \'000 US$)', fontweight='bold', fontsize=12)
    plt.grid(True, which='both', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

Graph Interpretations:

1. Yearly Economic Damages from Floods in Canada:

• While floods have been a consistent occurrence, their economic impact varies significantly over the years.
• The late 20th century and early 21st century showcase years with particularly high economic damages due to floods.

1. Yearly Economic Damages from Storms in Canada:

• Storms exhibit pronounced spikes in economic damages, especially in the latter half of the 20th century and early 21st century.
• The variability in economic damages underscores the unpredictable nature of storms and their potential for significant economic impact.

1. Yearly Economic Damages from Wildfires in Canada:

• Wildfires show an increasing trend in economic damages, especially from the late 20th century onwards.
• Some years in the 21st century have particularly high damages, emphasizing the growing economic challenge posed by wildfires.

1. Yearly Economic Damages from Extreme Temperature in Canada:

• Extreme temperature events, while frequent, generally result in lower economic damages compared to other disaster types.
• However, the late 20th century and early 21st century show a few years with pronounced economic damages due to extreme temperatures.

These insights highlight the varying economic challenges posed by different disaster types in Canada. While some disasters may occur frequently, their economic impact can differ significantly based on the severity and affected regions.

Next, we will proceed with the analysis of deaths over the years for each of the top 4 disaster types in Canada.

# Creating subplots for yearly deaths of the top 4 disaster types in Canada
plt.figure(figsize=(18, 12))

for i, disaster_type in enumerate(top_4_disasters_canada, 1):
    yearly_deaths_type = canada_data[canada_data['Disaster Type'] == disaster_type].groupby('Year')["Total Deaths"].sum()
    
    plt.subplot(2, 2, i)
    yearly_deaths_type.plot(color='crimson')
    plt.title(f'Yearly Deaths from {disaster_type} in Canada ({yearly_deaths_type.index.min()} - {yearly_deaths_type.index.max()})', fontweight='bold', fontsize=14)
    plt.xlabel('Year', fontweight='bold', fontsize=12)
    plt.ylabel('Number of Deaths', fontweight='bold', fontsize=12)
    plt.grid(True, which='both', linestyle='--', linewidth=0.5)

plt.tight_layout()
plt.show()

Graph Interpretations:

1. Yearly Deaths from Floods in Canada (1900 - 2021):

• Deaths due to floods in Canada vary over the years, with some years experiencing significant spikes.
• Despite floods being the most frequent disaster type, the death toll is not consistently high, indicating effective response measures in many instances.

1. Yearly Deaths from Storms in Canada (1900 - 2021):

• Storms have resulted in deaths throughout the timeframe, with some years having pronounced spikes.
• Similar to floods, while storms occur frequently, the death toll varies, emphasizing the unpredictable nature of such events.

1. Yearly Deaths from Wildfires in Canada (1900 - 2021):

• Wildfires have caused deaths sporadically over the years.
• The variability in the data suggests that while wildfires can be devastating, they do not consistently result in a high death toll.

1. Yearly Deaths from Extreme Temperature in Canada (1900 - 2021):

• Deaths due to extreme temperatures are sporadic but can be significant in certain years.
• The data underscores the potential for extreme temperature events to result in significant loss of life, even if they are not the most frequent disaster type.

With these insights, we have provided a comprehensive view of the challenges posed by natural disasters in Canada, both in terms of economic damages and the loss of life.

Insights Summary

With these analyses, we now have a comprehensive view of both global and Canada-specific insights regarding natural disasters. These insights are instrumental in understanding the scope, scale, and specific nuances of the challenges posed by natural disasters, thereby informing the design and features of our crisis response and management platform:

• Natural disasters have caused increasing economic damages globally, with pronounced spikes in the late 20th and early 21st centuries.
• Floods and storms are the most common disaster types globally and in Canada.
• The economic and human impact of natural disasters is highly variable, emphasizing the unpredictability and devastating potential of these events.
• Preparedness and rapid response mechanisms are crucial to mitigate the impacts of these disasters.

Conclusion

Understanding the patterns and impacts of natural disasters is essential for effective crisis response and management. With the insights derived from this analysis, we can make informed decisions, allocate resources more effectively, and design platforms and initiatives that address the most pressing challenges posed by natural disasters.