Health Monitoring and Analysis refers to the systematic use of health data to track and evaluate the health status of individuals or populations over time. It contains a range of activities from real-time physiological data collection (like heart rate, blood pressure, and temperature) to the analysis of more complex health records (including patient history, lifestyle choices, and genetic information). So, if you want to learn how to perform Health Monitoring and Analysis, this article is for you. In this article, I’ll take you through the task of Health Monitoring and Analysis using Python.
Health Monitoring and Analysis: Getting Started
The problem I am working on in this article is my take on a problem statement I found at statso. The dataset we are working with contains the following columns:
- PatientID: Numerical identifier for the patient.
- Age: Age of the patient in years.
- Gender: Gender of the patient.
- HeartRate: Heart rate in beats per minute.
- BloodPressure: Blood pressure readings, formatted inconsistently.
- RespiratoryRate: Respiratory rate in breaths per minute.
- BodyTemperature: Body temperature in Fahrenheit.
- ActivityLevel: Activity level at the time of the measurement.
- OxygenSaturation: Oxygen saturation percentage.
- SleepQuality: Quality of sleep reported by the patient.
- StressLevel: Reported level of stress.
- Timestamp: Date and time of the measurement.
You can download the dataset and read the problem statement here.
For the task of Health Monitoring and Analysis, we will aim to monitor the health of the patients in the data, analyze the patterns found in different types of patients and group them based on their health standards.
Health Monitoring and Analysis using Python
Now, let’s get started with the task of Health Monitoring and Analysis by importing the necessary Python libraries and the dataset:
import pandas as pd
health_data = pd.read_csv('healthmonitoring.csv')
print(health_data.head())PatientID Age Gender HeartRate BloodPressure RespiratoryRate \
0 1 69 Male 60.993428 130/85 15
1 2 32 Male 98.723471 120/80 23
2 3 78 Female 82.295377 130/85 13
3 4 38 Female 80.000000 111/78 19
4 5 41 Male 87.531693 120/80 14
BodyTemperature ActivityLevel OxygenSaturation SleepQuality StressLevel \
0 98.885236 resting 95.0 excellent low
1 98.281883 walking 97.0 good high
2 98.820286 resting 98.0 fair high
3 98.412594 running 98.0 poor moderate
4 99.369871 resting 98.0 good low
Timestamp
0 2024-04-26 17:28:55.286711
1 2024-04-26 17:23:55.286722
2 2024-04-26 17:18:55.286726
3 2024-04-26 17:13:55.286728
4 2024-04-26 17:08:55.286731
Let’s have a look at whether the data contains any null values or not:
health_data.isnull().sum()
PatientID 0
Age 0
Gender 0
HeartRate 0
BloodPressure 0
RespiratoryRate 0
BodyTemperature 18
ActivityLevel 0
OxygenSaturation 163
SleepQuality 0
StressLevel 0
Timestamp 0
dtype: int64
So, the data contains null values in body temperature and oxygen saturation columns. For simplicity, I’ll fill the null values using the median value:
# calculate medians median_body_temp = health_data['BodyTemperature'].median() median_oxygen_sat = health_data['OxygenSaturation'].median() # fill missing values health_data['BodyTemperature'].fillna(median_body_temp, inplace=True) health_data['OxygenSaturation'].fillna(median_oxygen_sat, inplace=True)
Next, we will examine summary statistics and the distribution of the numerical health metrics (Age, Heart Rate, Respiratory Rate, Body Temperature, and Oxygen Saturation). It will help us understand the typical values and the spread of the data. I’ll also include some visualizations to better understand these distributions:
import matplotlib.pyplot as plt
import seaborn as sns
sns.set(style="whitegrid")
# summary statistics
summary_stats = health_data.describe()
# plotting distributions of numerical features
fig, axes = plt.subplots(3, 2, figsize=(14, 18))
sns.histplot(health_data['Age'], bins=20, kde=True, ax=axes[0, 0])
axes[0, 0].set_title('Age Distribution')
sns.histplot(health_data['HeartRate'], bins=20, kde=True, ax=axes[0, 1])
axes[0, 1].set_title('Heart Rate Distribution')
sns.histplot(health_data['RespiratoryRate'], bins=20, kde=True, ax=axes[1, 0])
axes[1, 0].set_title('Respiratory Rate Distribution')
sns.histplot(health_data['BodyTemperature'], bins=20, kde=True, ax=axes[1, 1])
axes[1, 1].set_title('Body Temperature Distribution')
sns.histplot(health_data['OxygenSaturation'], bins=10, kde=True, ax=axes[2, 0])
axes[2, 0].set_title('Oxygen Saturation Distribution')
fig.delaxes(axes[2,1]) # remove unused subplot
plt.tight_layout()
plt.show()
print(summary_stats)
PatientID Age HeartRate RespiratoryRate BodyTemperature \
count 500.000000 500.000000 500.000000 500.000000 500.000000
mean 250.500000 51.146000 80.131613 17.524000 98.584383
std 144.481833 19.821566 9.606273 3.382352 0.461502
min 1.000000 18.000000 60.169259 12.000000 97.094895
25% 125.750000 34.000000 75.000000 15.000000 98.281793
50% 250.500000 51.000000 80.000000 17.500000 98.609167
75% 375.250000 69.000000 86.276413 20.000000 98.930497
max 500.000000 84.000000 99.925508 23.000000 99.489150
OxygenSaturation
count 500.000000
mean 96.296000
std 1.408671
min 94.000000
25% 96.000000
50% 96.000000
75% 97.000000
max 99.000000
Now, let’s have a look at the gender distribution in the data and the correlation between the numerical columns in the dataset:
# gender Distribution
gender_counts = health_data['Gender'].value_counts()
# correlation Matrix for numerical health metrics
correlation_matrix = health_data[['Age', 'HeartRate', 'RespiratoryRate', 'BodyTemperature', 'OxygenSaturation']].corr()
# plotting the findings
fig, axes = plt.subplots(1, 2, figsize=(12, 6))
# gender distribution plot
gender_counts.plot(kind='pie', ax=axes[0], autopct='%1.1f%%', startangle=90, colors=['#ff9999','#66b3ff'])
axes[0].set_ylabel('')
axes[0].set_title('Gender Distribution')
# correlation matrix plot
sns.heatmap(correlation_matrix, annot=True, fmt=".2f", cmap='coolwarm', ax=axes[1])
axes[1].set_title('Correlation Matrix')
plt.tight_layout()
plt.show()
The pie chart indicates a nearly even split between male and female subjects in the dataset, with males comprising a slight majority at 51.2%. The correlation matrix shows no strong correlations between the variables, as all the values are close to zero. Specifically, none of the health metrics (Age, Heart Rate, Respiratory Rate, Body Temperature, and Oxygen Saturation) display a strong positive or negative linear relationship with one another in this particular dataset. It suggests that, for this group of individuals, changes in one metric are not strongly associated with changes in the others.
Now, let’s have a look at the heart rate by activity level:
# heart Rate by activity level
plt.figure(figsize=(10, 6))
sns.boxplot(x='ActivityLevel', y='HeartRate', data=health_data)
plt.title('Heart Rate by Activity Level')
plt.ylabel('Heart Rate (beats per minute)')
plt.xlabel('Activity Level')
plt.show()
It shows that the median heart rate increases from resting to walking, which is expected as physical activity increases. However, the median heart rate does not significantly increase further during running compared to walking, which is unusual since we would expect a higher median heart rate for a more strenuous activity. Additionally, there is considerable overlap in the interquartile ranges between walking and running, suggesting similar variability in heart rates for these activities within the sampled population. The presence of outliers in the resting category indicates that some individuals have resting heart rates that are much higher than the typical range for the rest of the group.
Now, let’s have a look at the distribution of blood pressure levels and some health metrics by gender:
# extracting systolic and diastolic blood pressure for analysis
health_data[['SystolicBP', 'DiastolicBP']] = health_data['BloodPressure'].str.split('/', expand=True).astype(int)
# blood pressure distribution
plt.figure(figsize=(12, 6))
sns.histplot(health_data['SystolicBP'], color="skyblue", label="Systolic", kde=True)
sns.histplot(health_data['DiastolicBP'], color="red", label="Diastolic", kde=True)
plt.title('Blood Pressure Distribution')
plt.xlabel('Blood Pressure (mmHg)')
plt.legend()
plt.show()
# health metrics by gender
fig, axes = plt.subplots(1, 2, figsize=(14, 6))
sns.boxplot(x='Gender', y='HeartRate', data=health_data, ax=axes[0])
axes[0].set_title('Heart Rate by Gender')
axes[0].set_xlabel('Gender')
axes[0].set_ylabel('Heart Rate (beats per minute)')
sns.boxplot(x='Gender', y='OxygenSaturation', data=health_data, ax=axes[1])
axes[1].set_title('Oxygen Saturation by Gender')
axes[1].set_xlabel('Gender')
axes[1].set_ylabel('Oxygen Saturation (%)')
plt.tight_layout()
plt.show()
The systolic blood pressure, represented in blue, shows a more spread-out distribution with peaks suggesting common readings around 120 mmHg and 140 mmHg. The diastolic blood pressure, in red, appears to have a narrower distribution, with a significant peak around 80 mmHg. The spread of systolic values is broader than the diastolic ones, which is typical as systolic pressure tends to vary more with factors like activity level and stress. This distribution is consistent with general population trends where a systolic reading of around 120 mmHg and a diastolic reading of around 80 mmHg are considered normal.
For heart rate, both males and females show similar median values with a relatively similar interquartile range, indicating no significant difference in heart rate between genders within this dataset. In terms of oxygen saturation, again, both genders exhibit nearly identical medians and interquartile ranges, suggesting that oxygen saturation does not differ notably between males and females in this sample. There are a few outliers in oxygen saturation for both genders, indicating a few individuals with lower than typical values, but these do not seem to affect the overall distribution significantly.
Now, let’s analyze heart rate and oxygen saturation by sleep quality and stress levels:
# categorizing sleep quality and stress level for better analysis
sleep_quality_order = ['excellent', 'good', 'fair', 'poor']
stress_level_order = ['low', 'moderate', 'high']
# creating plots to examine relationships
fig, axes = plt.subplots(2, 2, figsize=(16, 12))
# heart rate by sleep quality
sns.boxplot(x='SleepQuality', y='HeartRate', data=health_data, order=sleep_quality_order, ax=axes[0, 0])
axes[0, 0].set_title('Heart Rate by Sleep Quality')
axes[0, 0].set_xlabel('Sleep Quality')
axes[0, 0].set_ylabel('Heart Rate (beats per minute)')
# heart rate by stress level
sns.boxplot(x='StressLevel', y='HeartRate', data=health_data, order=stress_level_order, ax=axes[0, 1])
axes[0, 1].set_title('Heart Rate by Stress Level')
axes[0, 1].set_xlabel('Stress Level')
axes[0, 1].set_ylabel('Heart Rate (beats per minute)')
# oxygen saturation by sleep quality
sns.boxplot(x='SleepQuality', y='OxygenSaturation', data=health_data, order=sleep_quality_order, ax=axes[1, 0])
axes[1, 0].set_title('Oxygen Saturation by Sleep Quality')
axes[1, 0].set_xlabel('Sleep Quality')
axes[1, 0].set_ylabel('Oxygen Saturation (%)')
# oxygen saturation by stress level
sns.boxplot(x='StressLevel', y='OxygenSaturation', data=health_data, order=stress_level_order, ax=axes[1, 1])
axes[1, 1].set_title('Oxygen Saturation by Stress Level')
axes[1, 1].set_xlabel('Stress Level')
axes[1, 1].set_ylabel('Oxygen Saturation (%)')
plt.tight_layout()
plt.show()
Heart rate appears relatively consistent across different levels of sleep quality and stress, with a slight increase in variation for those reporting poor sleep. Oxygen saturation shows a minimal decrease in median values from excellent to poor sleep quality, with some outliers indicating lower saturation for excellent and good sleep. When correlated with stress levels, oxygen saturation remains largely unchanged. Overall, while there are outliers, the central tendencies suggest that neither heart rate nor oxygen saturation is greatly affected by sleep quality or stress level within this dataset.
Now, let’s analyze the respiratory rate and body temperature by activity levels:
# creating plots to examine relationships between activity level and other health metrics
fig, axes = plt.subplots(1, 2, figsize=(16, 6))
# respiratory rate by activity level
sns.boxplot(x='ActivityLevel', y='RespiratoryRate', data=health_data, ax=axes[0])
axes[0].set_title('Respiratory Rate by Activity Level')
axes[0].set_xlabel('Activity Level')
axes[0].set_ylabel('Respiratory Rate (breaths per minute)')
# body temperature by activity level
sns.boxplot(x='ActivityLevel', y='BodyTemperature', data=health_data, ax=axes[1])
axes[1].set_title('Body Temperature by Activity Level')
axes[1].set_xlabel('Activity Level')
axes[1].set_ylabel('Body Temperature (°F)')
plt.tight_layout()
plt.show()
The respiratory rate tends to increase with activity level, as indicated by higher median rates for walking and running compared to resting. It aligns with physiological responses to exercise, where the breathing rate increases to meet oxygen demands. For body temperature, there is a slight upward trend from resting to running, which is consistent with the body heating up during physical exertion. There are outliers in body temperature at the resting and running levels, suggesting some individuals have temperatures outside the typical range for these activities. Overall, the trends observed are in line with expected physiological responses to varying levels of activity.
Grouping Patients
The data is not so complicated enough that we need to use a clustering algorithm to group patients. So, let’s group patients based on:
- Age Group: Young, Middle-aged, Senior
- Blood Pressure Category: Normal, Elevated, Hypertension Stage 1, Hypertension Stage 2
- Heart Rate Category: Low, Normal, High
- Oxygen Saturation Category: Normal, Low
# function to categorize Age
def age_group(age):
if age <= 35:
return 'Young'
elif age <= 55:
return 'Middle-aged'
else:
return 'Senior'
# function to categorize Blood Pressure
def bp_category(systolic, diastolic):
if systolic < 120 and diastolic < 80:
return 'Normal'
elif 120 <= systolic < 140 or 80 <= diastolic < 90:
return 'Elevated'
elif 140 <= systolic < 160 or 90 <= diastolic < 100:
return 'Hypertension Stage 1'
else:
return 'Hypertension Stage 2'
# function to categorize Heart Rate
def hr_category(hr):
if hr < 60:
return 'Low'
elif hr <= 100:
return 'Normal'
else:
return 'High'
# function to categorize Oxygen Saturation
def oxy_category(oxy):
if oxy < 94:
return 'Low'
else:
return 'Normal'
# applying categorizations
health_data['AgeGroup'] = health_data['Age'].apply(age_group)
health_data['BPCategory'] = health_data.apply(lambda x: bp_category(x['SystolicBP'], x['DiastolicBP']), axis=1)
health_data['HRCategory'] = health_data['HeartRate'].apply(hr_category)
health_data['OxyCategory'] = health_data['OxygenSaturation'].apply(oxy_category)
print(health_data[['Age', 'AgeGroup', 'SystolicBP', 'DiastolicBP', 'BPCategory', 'HeartRate', 'HRCategory', 'OxygenSaturation', 'OxyCategory']].head())Age AgeGroup SystolicBP DiastolicBP BPCategory HeartRate HRCategory \
0 69 Senior 130 85 Elevated 60.993428 Normal
1 32 Young 120 80 Elevated 98.723471 Normal
2 78 Senior 130 85 Elevated 82.295377 Normal
3 38 Middle-aged 111 78 Normal 80.000000 Normal
4 41 Middle-aged 120 80 Elevated 87.531693 Normal
OxygenSaturation OxyCategory
0 95.0 Normal
1 97.0 Normal
2 98.0 Normal
3 98.0 Normal
4 98.0 Normal
Now, let’s visualize the groups:
fig, axes = plt.subplots(2, 2, figsize=(16, 12))
# Age Group count plot
sns.countplot(x='AgeGroup', data=health_data, ax=axes[0, 0])
axes[0, 0].set_title('Distribution of Age Groups')
# Blood Pressure Category count plot
sns.countplot(x='BPCategory', data=health_data, ax=axes[0, 1])
axes[0, 1].set_title('Distribution of Blood Pressure Categories')
# Heart Rate Category count plot
sns.countplot(x='HRCategory', data=health_data, ax=axes[1, 0])
axes[1, 0].set_title('Distribution of Heart Rate Categories')
# Oxygen Saturation Category count plot
sns.countplot(x='OxyCategory', data=health_data, ax=axes[1, 1])
axes[1, 1].set_title('Distribution of Oxygen Saturation Categories')
# Show the plots
plt.tight_layout()
plt.show()
Observation:
- Distribution of Age Groups: The count plot shows that the ‘Senior’ category has the highest count, followed by the ‘Young’ and ‘Middle-aged’ categories. It suggests that seniors are the largest age group in this dataset.
- Distribution of Blood Pressure Categories: The majority of the dataset falls under ‘Normal’ blood pressure, with fewer instances in the ‘Elevated’ and ‘Hypertension Stage 1’. ‘Hypertension Stage 2’ has the lowest count, indicating that severe hypertension is less common among the participants.
- Distribution of Heart Rate Categories: Most individuals have a ‘Normal’ heart rate, with very few falling into the ‘Low’ or ‘High’ categories. It indicates that most individuals in this dataset have a heart rate that falls within the expected range.
- Distribution of Oxygen Saturation Categories: Almost everyone has ‘Normal’ oxygen saturation levels, with very few instances of ‘Low’ saturation. It suggests that oxygen deprivation is not a common issue in this group.
So, this is how you can perform Health Monitoring and Analysis using Python.
Summary
So, Health Monitoring and Analysis contains a range of activities from real-time physiological data collection (like heart rate, blood pressure, and temperature) to the analysis of more complex health records (including patient history, lifestyle choices, and genetic information).
I hope you liked this article on Health Monitoring and Analysis using Python. Feel free to ask valuable questions in the comments section below. You can follow me on Instagram for many more resources.
Thanks lot sir