This repository contains data and code for analyzing the results of an experiment conducted in Malawi to understand the impact of cash incentives on HIV disclosure. The project explores various factors, including age, distance to testing centers, and incentives received, to uncover insights into HIV disclosure behavior.
-
Incentive Impact: The majority of participants received incentives, and there is a moderate positive correlation between total incentives and HIV disclosure. This suggests that incentives may play a role in encouraging disclosure.
-
Demographic Influence: The age of participants varies widely, and there is an interesting relationship between age and incentive receipt. Further exploration is needed to understand the demographic factors influencing disclosure behavior.
-
Geographic Variances: The distance to testing centers varies across villages, and this may impact the likelihood of receiving incentives or learning HIV results. Tailored interventions may be needed for different geographic areas.
-
Model Predictions: The logistic regression model provides shows that incentives could play a role in enhancing HIV status disclosure.
- The dataset named
hiv_disclosure.csvwas used for the analysis. Kaggle dataset link: https://www.kaggle.com/datasets/utkarshx27/hiv-information-experiment
The dataset underwent a cleaning process to handle missing values and ensure a reliable analysis. The cleaned data includes relevant variables such as age, distance, total incentive, and HIV results. #Import neccessary libraries import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns
#import dataset hiv_disclosure=pd.read_csv("D:\hiv_disclosure.csv")
hiv_disclosure_cleaned = hiv_disclosure.dropna() Rename the "Unnamed: 0" column to "pt_id" hiv_disclosure_cleaned = hiv_disclosure_cleaned.rename(columns={'Unnamed: 0': 'pt_id'})
Basic descriptive statistics were calculated for key variables, providing insights into the age distribution, distance to testing centers, and total incentives. The summary includes mean, median, standard deviation, and other relevant measures.
descriptive_stats = hiv_disclosure_cleaned[['age', 'distvct', 'tinc']].describe() print("Descriptive Statistics:") print(descriptive_stats)
plt.figure(figsize=(10, 6)) sns.histplot(hiv_disclosure_cleaned['distvct'], bins=20, kde=True, color='salmon') plt.title('Distribution of Distance to Testing Center') plt.xlabel('Distance (km)') plt.ylabel('Frequency') plt.show()
plt.figure(figsize=(10, 6)) sns.histplot(hiv_disclosure_cleaned['tinc'], bins=20, kde=True, color='gold') plt.title('Distribution of Total Incentive') plt.xlabel('Total Incentive') plt.ylabel('Frequency') plt.show()
plt.figure(figsize=(10, 6))
sns.boxplot(x=hiv_disclosure_cleaned['age'], color='skyblue')
plt.title('Boxplot of Age')
plt.xlabel('Age')
plt.show()
plt.figure(figsize=(10, 6)) sns.boxplot(x=hiv_disclosure_cleaned['distvct'], color='salmon') plt.title('Boxplot of Distance to Testing Center') plt.xlabel('Distance (km)') plt.show()
plt.figure(figsize=(10, 6)) sns.boxplot(x=hiv_disclosure_cleaned['tinc'], color='gold') plt.title('Boxplot of Total Incentive') plt.xlabel('Total Incentive') plt.show()
percentage_received_incentive = ( hiv_disclosure_cleaned['any'].value_counts(normalize=True) * 100 )
print("Percentage of People Receiving Any Incentive:") print(percentage_received_incentive)
plt.figure(figsize=(8, 6)) sns.barplot(x=percentage_received_incentive.index, y=percentage_received_incentive.values, palette="pastel") plt.title('Percentage of People Receiving Any Incentive') plt.xlabel('Received Any Incentive') plt.ylabel('Percentage') plt.show()
plt.figure(figsize=(10, 6)) sns.boxplot(x='any', y='age', data=hiv_disclosure_cleaned, palette="pastel") plt.title('Relationship between Receiving Any Incentive and Age') plt.xlabel('Received Any Incentive') plt.ylabel('Age') plt.show()
plt.figure(figsize=(10, 6)) sns.boxplot(x='any', y='distvct', data=hiv_disclosure_cleaned, palette="pastel") plt.title('Relationship between Receiving Any Incentive and Distance to Testing Center') plt.xlabel('Received Any Incentive') plt.ylabel('Distance to Testing Center (km)') plt.show()
plt.figure(figsize=(10, 6)) sns.boxplot(x='got', y='distvct', data=hiv_disclosure_cleaned, palette="pastel") plt.title('Relationship between Learning HIV Results and Distance to Testing Center') plt.xlabel('Learned HIV Results') plt.ylabel('Distance to Testing Center (km)') plt.show()
plt.figure(figsize=(12, 8)) sns.boxplot(x='villnum', y='distvct', data=hiv_disclosure_cleaned, palette="pastel") plt.title('Distribution of Distances for Different Villages') plt.xlabel('Village ID') plt.ylabel('Distance to Testing Center (km)') plt.show()
plt.figure(figsize=(10, 6)) sns.boxplot(x='got', y='age', data=hiv_disclosure_cleaned, palette="pastel") plt.title('Relationship between Learning HIV Results and Age') plt.xlabel('Learned HIV Results') plt.ylabel('Age') plt.show()
plt.figure(figsize=(10, 6)) sns.countplot(x='got', hue='any', data=hiv_disclosure_cleaned, palette="pastel") plt.title('Relationship between Learning HIV Results and Receiving Any Incentive') plt.xlabel('Learned HIV Results (0: No, 1: Yes)') plt.ylabel('Count') plt.legend(title='Received Any Incentive', loc='upper right') plt.show()
correlation_matrix = hiv_disclosure_cleaned[['got', 'age', 'distvct', 'tinc', 'any']].corr()
plt.figure(figsize=(10, 8))
sns.heatmap(correlation_matrix, annot=True, cmap="coolwarm", fmt=".2f", linewidths=.5)
plt.title('Correlation Matrix')
plt.show()
The analysis revealed that approximately 78% of individuals received some form of incentive, while 22% did not.
Exploration of the relationship between receiving any incentive and other variables, such as age, was conducted. The percentage of people learning their HIV results based on whether they received an incentive was calculated.
The relationship between the distance to testing centers and the likelihood of receiving an incentive or learning HIV results was investigated. Visualizations were created to show the distribution of distances for different villages.
- Age: The average age of participants is 33.4 years, with a standard deviation of 13.6 years.
- Distance: The average distance to testing centers is 2.01 kilometers, with a standard deviation of 1.27 kilometers.
- Total Incentive: The average total incentive is 1.01, with a standard deviation of 0.90.
- Approximately 78% of individuals received some form of incentive, while 22% did not.
- The correlation coefficient between total incentive and HIV disclosure (got) is moderate at 0.35 while any and any(0.40).
- The other variables showed insignificant correlation coefficient values.
- The percentage of people learning their HIV results is higher among those who received incentives.
- Visualizations show varying distributions of distances for different villages.
-
The logistic regression model predicts HIV disclosure based on variables like age, distance, and total incentive.
-
Confusion Matrix: -[[ 75 95]
-
[ 49 346]] precision recall f1-score support
0.0 0.60 0.44 0.51 170 1.0 0.78 0.88 0.83 395accuracy 0.75 565 macro avg 0.69 0.66 0.67 565 weighted avg 0.73 0.75 0.73 565
-Precision for class 0 (did not learn HIV results) is 0.60, and recall is 0.44. -Precision for class 1 (learned HIV results) is 0.78, and recall is 0.88. -Overall accuracy is approximately 75%.
Contributions to this project are welcome. If you have ideas for improvements or additional analyses, please open an issue or submit a pull request.
This project is open-source and available under the MIT License. You are free to use, modify, and distribute the data and analysis results, but please give appropriate credit.