Introduction

The COVID-19 pandemic, caused by the novel coronavirus SARS-CoV-2, has rapidly
evolved into one of the most significant global health crises of the 21st century,
profoundly impacting individuals, communities, and nations across the globe. First
identified in December 2019 in Wuhan, China, the virus quickly spread across borders,
leading to widespread transmission and significant morbidity and mortality. The World
Health Organization (WHO) declared COVID-19 a Public Health Emergency of
International Concern in January 2020, followed by a declaration of a pandemic in March
2020, as the virus continued its relentless spread to virtually every corner of the world.
Since then, efforts to contain the virus, mitigate its impact, and develop effective
interventions have been at the forefront of global health agendas, with researchers,
healthcare professionals, policymakers, and communities working tirelessly to confront
this unprecedented challenge.

The emergence of SARS-CoV-2, a novel beta coronavirus, has presented unique

challenges due to its high transmissibility, variable clinical presentation, and potential for
severe illness and death, particularly among vulnerable populations. While respiratory
transmission remains the primary mode of spread, the virus has demonstrated the ability
to infect multiple organ systems, leading to a wide spectrum of clinical manifestations
ranging from mild flu-like symptoms to severe pneumonia, acute respiratory distress
syndrome (ARDS), and multi-organ failure. Furthermore, the evolving nature of the
pandemic, characterized by the emergence of new variants and the uneven distribution of
vaccines and resources, underscores the need for ongoing research, surveillance, and
collaboration to navigate the complexities of the COVID-19 landscape. In light of the
multifaceted challenges posed by COVID-19, this paper aims to provide a comprehensive
overview of the current state of knowledge surrounding the virus, its epidemiology,

1
clinical characteristics, public health implications, and socio-economic impacts. By
synthesizing findings from diverse research disciplines, we seek to contribute to a deeper
understanding of the pandemic and inform evidence-based strategies for addressing its
immediate and long-term consequences. We can harness the collective expertise and
resources needed to overcome this unprecedented global health crisis and build a more
resilient future for all through interdisciplinary collaboration and global solidarity.

Abstract
The project presents an in-depth analysis of the COVID-19 pandemic, caused by the
SARS-CoV-2 virus, highlighting its extensive impacts on public health, economies, and
societies worldwide. Epidemiological studies are crucial for understanding the
distribution and determinants of the disease, utilizing data from national health databases
and public health reports to analyze trends, identify hotspots, and explore factors
influencing virus spread through statistical tools and regression analysis. These studies
are vital for identifying high-risk areas and populations for targeted interventions.

Surveys on public health behavior are another key method, gathering data on individual
behaviors, attitudes, and perceptions through online surveys of 1,000 participants.
Descriptive and inferential statistics help summarize the data and explore associations
between demographics and health behaviors, providing insights into public adherence to
guidelines and informing health campaigns and policies.

A mixed-methods approach enriches the analysis by combining qualitative data from

in-depth interviews with healthcare workers and patients with quantitative survey data
from the general public. This approach uses thematic analysis for qualitative data and
statistical software for quantitative data, offering a comprehensive understanding of

2
COVID-19's impact. The synthesis of findings across virological, epidemiological,
clinical, and socioeconomic research underscores the multifaceted nature of the
pandemic. Virological studies have detailed the virus's genomic structure and
transmission dynamics, aiding in diagnosis and vaccine development. Clinical research
has clarified COVID-19's clinical spectrum, while public health studies have assessed
intervention effectiveness. Socio-economic research has highlighted the pandemic's
broader impacts, emphasizing the need for sustained research, interdisciplinary
collaboration, and global solidarity to address ongoing challenges and prepare for future
pandemics.

Methodology
1. Study Design
This study employs a retrospective cohort study design to investigate the impact of
COVID-19 on health outcomes and healthcare utilization among individuals with
underlying health conditions. The retrospective design allows for the examination of
historical data from electronic health records (EHRs) to assess outcomes over time.

2. Study Setting
The study is conducted at a tertiary care medical center serving a diverse patient
population in Mexico. The institution has comprehensive EHR systems capturing detailed
patient information, including demographic data, clinical diagnoses, laboratory results,
and treatment records.

3. Study Population
The study population comprises adult patients (aged 18 years and above) with
documented underlying health conditions, including but not limited to hypertension,

3
diabetes, cardiovascular disease, chronic respiratory conditions, and
immuno-compromised states. Patients with confirmed or suspected COVID-19 infection
are identified based on laboratory testing, diagnostic codes, or clinical documentation.

4. Data Collection
Data for the study are extracted from the institution's EHR database. Relevant variables
extracted include demographic characteristics (age, sex), underlying health conditions,
COVID-19 diagnosis (laboratory-confirmed or clinically suspected), disease severity
(mild, moderate, severe), treatment modalities (pharmacological and
non-pharmacological), healthcare utilization (hospital admissions, emergency department
visits, intensive care unit admissions), and clinical outcomes (mortality, morbidity, length
of hospital stay).

5. Data Analysis
Descriptive statistics are used to summarize the demographic and clinical characteristics
of the study population. Bivariate analysis, including chi-square tests and t-tests, is
conducted to compare outcomes between patients with and without COVID-19 infection
and across different subgroups based on disease severity and comorbidity profiles.
Multivariable regression analysis, such as logistic regression or Cox proportional hazards
models is employed to assess the association between COVID-19 infection and health
outcomes while controlling for potential confounding variables (e.g., age, sex,
comorbidities).

6. Ethical Considerations
The study protocol has been reviewed and approved by the Institutional Review Board
(IRB) . Patient confidentiality and privacy are maintained throughout the study, with data
anonymized and stored securely in compliance with relevant regulatory guidelines.

4
Justification

The COVID-19 pandemic has undoubtedly emerged as one of the most challenging
global crises in recent history, impacting virtually every aspect of human life. This project
undertakes a comprehensive analysis of the pandemic, examining its epidemiological,
socio-economic, healthcare, and future outlook dimensions.

At its core, the COVID-19 pandemic is a health crisis characterized by the rapid spread of
a novel coronavirus, SARS-CoV-2. The virus has demonstrated formidable
transmissibility, resulting in a global pandemic within a matter of months. Understanding
the epidemiology of COVID-19 is crucial for grasping the scale and trajectory of the
crisis. This includes studying transmission dynamics, the emergence of variants, and the
disproportionate impact on vulnerable populations such as the elderly and those with
underlying health conditions.

However, the ramifications of COVID-19 extend far beyond the realm of public health.
The pandemic has triggered a cascade of socio-economic consequences, disrupting
economies, livelihoods, and social structures worldwide. Lockdown measures aimed at
curbing the spread of the virus have led to economic recession, widespread job losses,
and exacerbation of existing inequalities. Moreover, disparities in access to healthcare
services have been magnified, with marginalized communities bearing the brunt of the
pandemic's impact.

In response to the crisis, governments, healthcare systems, and communities have

implemented various strategies to mitigate the spread of the virus and alleviate its effects.
Public health measures such as social distancing, mask mandates, and vaccination
campaigns have played a critical role in controlling transmission. However, challenges
such as vaccine hesitancy, supply chain disruptions, and inequitable distribution have

5
complicated efforts to achieve widespread immunity. On the healthcare front, innovations
in treatment, telemedicine, and healthcare delivery have emerged, reshaping the
landscape of medical practice amid a pandemic.

Looking ahead, the future trajectory of COVID-19 remains uncertain, shaped by factors
such as vaccination coverage, the emergence of new variants, and societal responses.
While vaccination offers hope for controlling the spread of the virus and returning to
pre-pandemic normalcy, ongoing vigilance and preparedness are essential to navigate
potential challenges. Long-term consequences of the pandemic, including its impact on
mental health, education, and global governance, will require sustained attention and
resources.

The COVID-19 pandemic represents a watershed moment in human history, underscoring

the interconnectedness of global health, economies, and societies. By undertaking a
comprehensive analysis of the pandemic, this project aims to deepen our understanding of
its multifaceted impact and inform evidence-based responses for building more resilient
and equitable societies in the post-pandemic era. As we continue to navigate the
complexities of COVID-19, solidarity, cooperation, and science-based decision-making
will be indispensable in overcoming this unprecedented challenge.

6
Data Analysis & Visualization

a. Heatmap to show the correlation between all variables

Fig 1: Heatmap portraying the correlation between various variables

7
According to Fig 1: Heatmap portraying the correlation between various variables,
we get the following analysis:

1. High Correlations:
Some pairs of variables show high positive correlations (values close to 1). For example,
age and the elderly correlate by 1.00, indicating that these two variables are perfectly
correlated. Similarly, ICU and intubated have a high correlation, suggesting that patients
in the ICU are often intubated.

2. Negative Correlations:
Some pairs of variables show high negative correlations (values close to -1). For instance,
sex and pregnancy have a correlation of -1.00, which makes sense biologically as
typically only females can be pregnant.

3. Low or No Correlations:
Many variables show low or near-zero correlations, indicating weak or no linear
relationship between those variables. For instance, sex and age show a correlation of
0.03, suggesting there is no significant linear relationship between these two variables.

4. Distinct Clusters:
Variables such as diabetes, COPD, asthma, hypertension, and other diseases cluster
together, indicating they might be related or occur together frequently in patients.

5. Anomalies:
Some correlations might appear unexpected and could warrant further investigation. For
example, intubation and pneumonia have a correlation of 0.46, which might indicate a
moderate relationship but should be explored further to understand the context better.

8
 b. Histogram to show the age distribution

Fig 2: Histogram to show the age distribution

According to Fig 2: Histogram to show age distribution, we get the following analysis:

● This age distribution suggests a dataset heavily represented by middle-aged adults,

with additional significant groups of young children and fewer older individuals.

9
 c. Heatmap representing the correlation coefficients between pairs
of variables.

Fig 3: Heatmap representing the correlation coefficient between pairs of variables

According to Fig 3: Heatmap representing the correlation coefficient between pairs

of variables, we get the following analysis:

● There is a strong correlation between obesity and chronic diseases.

● Elderly patients tend to be hospitalized.
● Elderly patients tend to die.
● Patients tend to have multiple diseases.
● Patients with chronic diseases are immunosuppressed

10
 d. Count plot generated using the Seaborn library

Fig 4: Count plot generated using the Seaborn library

According to Fig 4: Count plot generated using the Seaborn library visualizes the
number of patients (represented by count on the y-axis) across different medical units
(represented by medical_unit on the x-axis), categorized by death status. The hue
parameter in the plot differentiates the counts by death status: '1' likely represents
survivors and '2' represents deaths.

● Unit 4 has the highest number of deaths: This is evident from the significant height
of the black bar (death = 2) for medical_unit 4.

● Unit 12 has the highest overall patient count: This is indicated by the tallest bars
(both for survivors and deaths) compared to other units.

11
 ● Other units: Units 1, 2, 3, 5, 6, 7, 8, 9, 10, 11, and 13 have comparatively lower
counts of patients and deaths.

The text below the plot incorrectly states "Unit 4 have the highest percentage of death,"
which should be corrected to reflect that Unit 4 has the highest number of deaths, not
necessarily the highest percentage.

12
 e. Bar plot created with the Pandas plotting functionality in
Python, visualizing the number of deaths per month

Fig 5: Barplot visualizing the number of deaths per month

According to Fig 5: Barplot visualizing the number of deaths per month visualizes
the number of patients (represented by count on the y-axis) across different medical units
(represented by medical_unit on the x-axis), categorized by death status. The hue
parameter in the plot differentiates the counts by death status: '1' likely represents
survivors and '2' represents deaths.
● Highest Deaths in June (Month 6): The month of June shows the highest number of
deaths, reaching close to 25,000.
● Significant Deaths in May and July (Months 5 and 7): Both May and July also show
high death counts, though slightly lower than June, with counts between 20,000 and
25,000.
● Moderate Deaths in April and August (Months 4 and 8): April and August have

13
 moderate death counts, ranging between 15,000 and 20,000.
● Low Deaths in Other Months: January, February, March, September, October,
November, and December show significantly lower death counts, all below 5,000.
This plot helps identify the peak months for deaths, which might be useful for
resource planning and understanding seasonal variations in mortality.

14
 f. Bar chart to show the percentage of death for each factor

Fig 6: Barchart to show the percentage of death for each factor

According to Fig 6: Barchart to show the percentage of death for each factor, we get
the following analysis:

● Pregnancy has the lowest percentage of death among the factors listed, indicating
that death incidence in pregnant women due to COVID-19 is low.
● Pneumonia has the highest percentage at 39%, followed by the elderly at 32% and
renal chronic conditions at 30%.
● Other significant factors include COPD (27%), diabetes (23%), cardiovascular
diseases (21%), and hypertension (20%).
● Conditions like immunosuppression and obesity also contribute to the death
percentage but are not as high as the aforementioned factors.
● Asthma and tobacco use have relatively lower percentages of death.

15
 g. The histogram illustrates the distribution of ages among
pregnant women in the dataset.

Fig 7: Histogram illustrating the distribution of ages among pregnant women

According to Fig 7: Histogram illustrating the distribution of ages among pregnant

women, we get the following analysis:

● The chart indicates that the majority of pregnant women in the dataset are between 18
and 35 years old.
● There are few cases of pregnant women above 40, and the frequency drops significantly
for ages beyond 35.
● The dataset contains very few instances of pregnant women aged below 18 and above
35, indicating that pregnancies in these age ranges are relatively rare in the context of
this data.
● The data suggests a concentration of pregnant women within a typical childbearing age
range, aligning with general demographic trends.

16
 PYTHON CODES FOR

VISUALISATION AND ANALYSIS

17
 import os

for dirname, _, filenames in os.walk('/kaggle/input'):

for filename in filenames:

print(os.path.join(dirname, filename))

adress="/kaggle/input/covid19-dataset/Covid Data.csv"

#Importing

import pandas as pd

import seaborn as sns

import matplotlib.pyplot as plt

%matplotlib inline

import plotly.express as px

import datetime

# Load of data

df = pd.read_csv(adress)

18
 Exploratory And Cleaning of Data

print("SHAPE OF DATAFRAME is: ", df.shape)

df.head()

df.isnull().sum().sum()

df.info()

#lower name of columns

df.rename(columns=lambda x : x.lower(), inplace=True)

#Parsing of date

df['date_parsed']=pd.to_datetime(df['date_died'][df['date_died']!="9999-99-99"]

, format="%d/%m/%Y")

df['date_died'].value_counts()

#convert date_died to died or not, 2 foe alive 1 for dead

df['death'] = [2 if each == "9999-99-99" else 1 for each in df['date_died'] ]

#drop date_died column

df.drop(columns="date_died", inplace=True)

#Number of unique values of columns

df.nunique()

19
 #Create a column for ptients > 65 years and < 65 years

df['elderly'] = [2 if each < 65 else 1 for each in df['age']]

df['elderly'].value_counts()

#calculation of percentage of null values

for i in df.loc[:, ~df.columns.isin(["date_parsed","age", "medical_unit",

"clasiffication_final"])]:

null = df[i][(df[i]!=1)& (df[i]!=2)].value_counts()

if null.any():

pe = (null/1048575)*100 #where 1048575 is length of column

print("percentage of null values of {} are: \n{}".format(i,pe))

for i in df.loc[:, ~df.columns.isin(["date_parsed", "age"])]:

f = sns.countplot(x=df[i]);

plt.show()

df[(df['pregnant']==97)].sex.value_counts()

#Convert 97 value to 2 in pregnant column men (men do not get pregnant)

df['pregnant'].replace(97, 2, inplace=True)

plt.figure(figsize=(16,16))

sns.heatmap(data=df.corr(), annot=True, fmt=".2f");

20
#Drope columns which do not have correlation or helpless and contain many nan
values

col = ["sex", "intubed", "clasiffication_final", "icu"]

df.drop(columns = col, inplace=True)

df.shape

sns.histplot(data=df['age']);

#correlation of data

plt.figure(figsize=(16, 6))

sns.heatmap(df.corr(), annot=True);

# death acc to medical_unite

sns.countplot(hue=df.death, x= df.medical_unit);

#Creat column for month

df['month']=df["date_parsed"].dt.month

df.groupby("month")["death"].count().plot(kind="bar");

plt.title("Number of Deaths per month");

21
#Function to calculate percentage of death depend on charachteristic factors of
patiens

"""

get total number of patients with some charachreristic and died number of same
charach

and get percentage of died patient

"""

def perc_die(col_name,has=1, die=1):

total=df[df[col_name]==has].age.count()

num_died=df[(df[col_name]==has) & (df['death']==die)].age.count()

return num_died/total*100

#Calculate total percentage of death

t = df[df['death']==1].death.count()/df.shape[0]*100

print("total percentage of sample death = {}%".format("%.2f"%t))

#Calculation of death percentage to detect the effect of each on incidence of

death

percen = []

charc_cols=["inmsupr", "pregnant", "asthma","hipertension", "elderly",

"obesity", "cardiovascular", "renal_chronic", "tobacco", "other_disease",

22
 "copd", "diabetes", "pneumonia"]

for i in charc_cols:

p = perc_die(i)

print("Percentage of death of patients with {} is: {}%".format(i,

"%.2f"%p))

percen.append(p)

print("Maximum percentage of death is: {}".format("%.2f"% max(percen)))

px.bar(x=charc_cols, y=percen, text_auto=".2s", title="Percentage of death for

each factor(reflect effect of each on death):")

#Function to calculate percentage of aquired pneumonia depend on charachteristic

factors of patiens

"""

get total number of patients with some charachreristic and pneumonia patients
number of same charach

and get percentage of died patient

"""

def perc_pnm(col_name,has=1, pneum=1):

total=df[df[col_name]==has].age.count()

num_died=df[(df[col_name]==has) & (df['pneumonia']==pneum)].age.count()

23
 return num_died/total*100

#Calculate total percentage of patients who aquired pneumonia

t = df[df['pneumonia']==1].death.count()/df.shape[0]*100

print("total percentage of pneumonia = {}%".format("%.2f"%t))

#Calculationalign of pnenmenia percentage

pnm=[]

for i in charc_cols[:-1]:

pr = perc_pnm(i)

print("Percentage of pneumonia of patients with {} is: {}%".format(i,

"%.2f"%pr))

pnm.append(pr)

px.bar(x=charc_cols[:-1], y=pnm, text_auto=".2s", title="Percentage of patients

which acquired pneumnia acc to disease type:")

df.columns

df[(df['death']==2)&(df['pregnant']==1)].age.max()

child=df[(df['age']<16)&

(df["death"]==1)].age.count()/df[df['age']<16].age.count()*100

print("percentage of children died from covid = {}%".format("%.2f"%child))

24
 ages=df[df['pregnant']==1].age.unique()

print("Age of pregnant women: ", ages)

plt.title("Age of pregnant women")

df[df['pregnant']==1].age.hist();

ML Model To Predict Death

# import helpful libraries

from sklearn.tree import DecisionTreeRegressor

from sklearn.metrics import mean_absolute_error

from sklearn.model_selection import train_test_split

df.columns

features=["inmsupr","hipertension", "elderly","cardiovascular",

"renal_chronic", "other_disease", "copd", "diabetes", "obesity", "tobacco",

"age", "pneumonia"]

#Separate target data

y = df.death.iloc[:-10]

x= df.loc[:,features].iloc[:-10]

new_data= df.loc[:, features].iloc[-10:]

result_nw_data=df['death'].iloc[-10:]

25
 x.head()

sns.countplot(x=y); # that will make unacurate prediction

#samble size

df.shape

# Resambeling of data

from imblearn.over_sampling import RandomOverSampler

ros = RandomOverSampler(random_state=0)

x_re,y_re = ros.fit_resample(x,y)

sns.countplot(x=y_re);

#Split data to validation and training

train_x, val_x, train_y, val_y = train_test_split(x_re, y_re, random_state=1)

#Define dcision tree model

model = DecisionTreeRegressor(random_state=1)

model.fit(train_x, train_y)

prediction = model.predict(val_x)

prediction

26
 #Make dataframe to compare prediction with real data

real_pred={"Real":list(val_y[10:20]),

"Predict": list(prediction[10:20])}

df_real_pred=pd.DataFrame(real_pred)

df_real_pred

#Evaluate model

mae = mean_absolute_error(prediction, val_y)

print("Validation MAE for Desicion Tree Model: {:,f}".format(mae))

#test of model by new data

list(model.predict(new_data))

# check real values

result_nw_data

27
Limitations
Several limitations are acknowledged in this study, including the retrospective study
design, reliance on secondary data sources, potential for selection bias, and limitations
inherent to observational studies, such as unmeasured confounding and residual bias.

The Limitations of the study are as follows:

● Data Quality and Reporting Bias: Studies on COVID-19 may be limited by
inconsistencies in data reporting, variations in testing protocols, and underreporting or
misclassification of cases. This can lead to inaccuracies in estimating disease burden,
transmission dynamics, and outcomes.
● Sampling Bias: Research on COVID-19 may suffer from sampling bias if the study
population is not representative of the broader population. For example, studies
conducted in healthcare settings may overrepresent severe cases, while community-based
studies may miss asymptomatic or mild cases.
● Limited Longitudinal Data: Many studies on COVID-19 are cross-sectional or have
short follow-up periods, limiting their ability to assess long-term outcomes, such as the
persistence of symptoms (long COVID), the duration of immunity after infection or
vaccination, and the efficacy of treatments over time.
● Confounding Factors: Studies examining the association between COVID-19 outcomes
and various factors (e.g., comorbidities, demographics, interventions) may be confounded
by unmeasured or inadequately controlled variables. Failure to account for confounders
can lead to biased risk factors or treatment effects estimates.
● Variability in Testing and Diagnostic Criteria: Differences in testing availability,
diagnostic criteria, and healthcare-seeking behavior across regions and periods can affect
the comparability of study results and the accuracy of prevalence and incidence estimates.

28
● Evolution of the Virus: The ongoing evolution of SARS-CoV-2 and the emergence of
new variants may impact the generalizability and relevance of research findings over
time. Studies conducted early in the pandemic may not capture the dynamics of more
recent viral strains or the effectiveness of vaccines against them.
● Resource Constraints and Ethical Considerations: Resource constraints, including
limited funding, personnel, and infrastructure, may affect the design, conduct, and
interpretation of COVID-19 research. Ethical considerations, such as patient privacy,
informed consent, and equity in access to interventions, may also pose challenges to
study implementation.
● Publication Bias and Preprint Influence: Studies with positive or significant results may
be more likely to be published or disseminated, leading to publication bias and potentially
inflating the perceived effectiveness of interventions or associations. Preprint servers and
rapid dissemination channels may amplify the influence of preliminary findings before
peer review.

29
Conclusion

The COVID-19 pandemic has emerged as a defining global health crisis of the 21st
century, presenting unprecedented challenges to societies, economies, and healthcare
systems worldwide. Through a comprehensive review of the literature, this paper has
synthesized key findings and insights from research across various domains, including
virology, epidemiology, clinical management, public health interventions, and
socio-economic impacts.

Despite significant progress in understanding and combating COVID-19, challenges

remain as the pandemic continues to evolve. The emergence of new variants, disparities
in vaccine distribution, and the enduring socio-economic consequences underscore the
importance of ongoing research, collaboration, and innovation in addressing the complex
and interconnected challenges posed by the pandemic. Moving forward, it is imperative
to maintain vigilance and adaptability in our response to COVID-19, leveraging
evidence-based strategies, interdisciplinary collaboration, and global solidarity.
Investments in public health infrastructure, healthcare systems, and social support
mechanisms are essential to mitigate the immediate crisis and build resilience for future
pandemics. As we navigate the uncertainties and complexities of the COVID-19
landscape, we must remain committed to equity, solidarity, and compassion, recognizing
that our collective efforts will shape the trajectory of the pandemic and determine the
health and well-being of generations to come. By working together, we can overcome the
challenges posed by COVID-19 and build a more resilient and equitable world for all.

30
References

● Referred to Kaggle on 16/02/2024 to collect dataset

https://datasetsearch.research.google.com/search?src=3&query=
covid%2019&docid=L2cvMTFqel93cDY1cg%3D%D

● Referred to Datos on 23/02/2024

https://datos.gob.mx/busca/dataset/informacion-referente-a-caso
s-covid-19-en-mexico

● Referred to World Health Organisation on 4/03/2024

https://www.who.int/emergencies/diseases/novel-coronavirus-20
19/situation-reports

https://www.worldometers.info/coronavirus/

● Referred to Covid 19 sample collection management system

https://covid19cc.nic.in/icmr/Citizen/TestSampleReport.aspx

31