Tackling Health Crisis In Africa

I decided to present a solution on a hackathon organized earlier this year by 10Alytics, a data and strategy company, and the hackathon was focused on the theme “Tackling the health crisis in Africa”. My contribution to the project involved conducting a comprehensive analysis to derive insights and findings on the factors affecting the health crisis in Africa. The aim was to provide the right solutions and recommendations to address the crisis.

To carry out the analysis, I utilized Microsoft Power BI for the entire project, including data cleaning, modeling, analysis, and visualization. The dashboard images below showcase the results of my analysis. You can also interact with the dashboard by clicking the link provided:

Power BI Report

N.B: Please note that all visuals related to the Crude Mortality Rate and Medical Personnel are measured per 100,000 population. Specifically, the Crude Mortality Rate represents the number of deaths per 100,000 people, while Medical Personnel refers to the number of healthcare workers per 100,000 of the country’s population.

Dataset Overview

There were six datasets provided for this project. They are as follows:
1. Causes of Deaths : This indicates the number of deaths and the causes across different countries from 1990 to 2019
2. Number of Deaths by Age Group: This indicates the number of deaths by Age Group across different countries from 1990 to 2019.
3. Medical Doctors per 10,000 population : This indicates the number of medical doctors per 10,000 population in Africa.
4. ISO 3166_Country-and-continent-codes-list-csv: This contains the list of sovereign states and dependent territories by continent describing the list of countries by continent.
5. World Population: This indicates the data from United Nation Population Division across countries over the last century.
6. Health Expenditure (% of GDP) : This data contains countries expenditure on health as a percentage of GDP

Data Preparation

I imported all the datasets into power query, where I cleaned all the data. These were common steps I took in cleaning each of the dataset.

1. As some datasets contained information on countries outside of Africa, I filtered out those rows and kept only the rows that pertain to African countries. This was done to avoid redundancy in the data.
2. After filtering out non-African countries, I noticed that some entries in the numerical datatype column contained null values. Since these null values are known to represent empty data, I replaced them with 0, rather than removing those rows entirely and losing valuable data.
3. To ensure clarity when creating measures, I renamed necessary columns.
4. I removed duplicates from all the tables.
5. I filtered the tables to display data from the years 2000 to 2019, as this time range is the most consistently shared range among all the tables.
6. I merged the tables with the “ISO 3166_Country-and-continent-codes-list-csv” table on the 3-letter country code- (A column that contains unique code for each country name) in order to have only African countries in the tables and to ensure the tables have the same country names. This step will also help in the data modeling because the “ISO 3166_Country-and-continent-codes-list-csv” will server as a dimension table as it contains all African unique country names and codes.
7. To avoid having excessively long table names that occupy too much space when creating DAX measures, I renamed some of the tables. For instance, the original name of the “ISO 3166_Country-and-continent-codes-list-csv” table seemed too long, so I changed it to “Dim_Country”.

N.B: These are few cleaning steps common in each of the tables, These are not entirely all the steps I took in cleaning the data.

DATA MODELING

I created a calendar table in Power Query that ranges from the year 2000 to 2019, as I mentioned earlier. I used this year range because it is the most consistent range in each of the tables. I would like to have a date/calendar table and a dimension country table that will have a one-to-many relationship with the remaining tables, as all the remaining tables are fact tables and they all have a date column and a country code. Therefore, when writing DAX measures and creating visualizations, I can use a table from which to pull date columns and country columns. Below is a screenshot of the data model, which has a star schema structure. The screenshot below shows the data model in detail.

FINDINGS

1. Lesotho, Eritrea, and Somalia had the highest crude mortality rates among all countries, with approximately 1437, 1217, and 1035 deaths per 100,000 population, respectively. In contrast, Mauritania, Algeria, and Sudan had the lowest crude mortality rates, with approximately 460, 450, and 424 deaths per 100,000 population, respectively.

2. I observed that Sierra Leone, Namibia, South Africa, Burundi, and Lesotho were the top five countries that spent the highest percentage of their GDP on healthcare. Interestingly, none of these countries were among the countries with the lowest mortality rates. In fact, Lesotho, which was one of the top countries for healthcare spending, had the highest mortality rate.

3. The top 5 Countries with the highest population are Nigeria, Ethiopia, Egypt, Tanzania and Kenya.

4. After exploring the crude mortality rate by age group, it was found that the “Under 5” age group had the highest rate. This may be attributed to the weaker immune systems in young children, which make them more susceptible to neonatal disorders, lower respiratory infections, and diarrheal diseases.

5. Cardiovascular diseases, neonatal disorders, and lower respiratory infections are the three leading causes of death, with crude mortality rates of approximately 145, 73, and 65 deaths per 100,000 population, respectively.

6. Out of the countries included in the study, Mauritius, Seychelles, and Algeria had the highest number of medical doctors per 100,000 population, with 2806, 1463, and 603 doctors respectively. On the other hand, Ethiopia, Tanzania, and Liberia had the lowest number of medical doctors per 100,000 population, with only 20, 14, and 15 doctors respectively.

KEY INSIGHTS

1. The crude mortality rate has steadily declined from 2000 to 2019, and this indicates an increase in preventive measures to reduce mortality.
2. Between 2000 and 2019, there was a gradual decrease in crude mortality rates across all age groups.
3. The African population has steadily increased over the years, reaching its highest recorded level of 1.11 billion in 2019.
4. The trend of health GDP was erratic, with the highest growth rate occurring in 2009. However, it experienced a slow decline from 2015 until the end of the year.
5. . Between 2001 and 2002, medical personnel saw a massive increase in numbers, with a growth rate of 2202.24% compared to the previous year.
   . The trend continued in an irregular manner until another sharp increase of 779.4% was observed between 2015 and 2016.
   . However, between 2018 and 2019, there was a significant decline in the number of medical personnel, with a growth rate of -79.29% compared to the previous year.

RECOMMENDATIONS

1. We observed that children under the age of five have the highest mortality rates, often due to inadequate healthcare. It’s clear that we need more pediatricians to take care of these children, especially in rural areas where diseases are more prevalent. We urge the government to provide more free healthcare services to these areas and raise awareness among mothers about preventative measures they can take to protect their children from contracting these diseases. By investing in these measures, we can significantly reduce the suffering and mortality rates of young children.
2. Countries with high mortality rates such as Lesotho, Eritrea and Somalia need to invest more in their healthcare and healthcare facilities, increasing the number of medical professionals. Also, support from international organizations such as the World Health Organization (WHO) can be requested.
3. While healthcare spending serve as a major factor, it’s not always directly related to lower mortality rates. Countries like Lesotho, as we observed earlier spend a high percentage of their GDP on health care, but still experience high mortality rates, should reevaluate their health care systems to detect inefficiencies and areas that require support.
4. Countries with low numbers of medical doctors such as Ethiopia, Tanzania, and Liberia should invest in training and educationg more medical professionals to help this shortage.
   Measures such as offering incentives to encourage more individuals study medicine can be implemented, and international organizations can provide support in terms of resources and training programs.
5. Efforts to reduce the mortality rates such as vaccination programs, seminars to increase awareness of risk factors should be focused on addressing the leading causes of death such as cardiovascular diseases, neonatal disorders and lower respirator infections.
6. Top populous countries such as Nigeria, Ethiopia, Egypt, Tanzania and Kenya should review their health systems to ensure they are capable of meeting the demands of their population.

I thoroughly enjoyed working on this challenging project, particularly when writing complex DAX measures. It was all worth it in the end, as we were able to generate valuable insights. Once again, You can interact with the dashboard here. I’d like to extend my appreciation to 10alytics for hosting this exciting hackathon.
Also, If you enjoyed this article and would like to share suggestions, you can connect with me on Linkedin and Twitter 😅. You can also check out my other projects on my github.