Improving the Life Quality of Kids through Analytics

Comments · 312 Views

Each year, UNICEF’s flagship publication, The State of the World's Children, closely examines a key issue affecting children. Several statistics for categories such as Basic Indicators, Nutrition, Health, HIV/AIDS, Education, Demographic Indicators, etc. have been provided. I have analyzed this data and come up with a few recommendations to improve the quality of life of children.

In a world where inequality is ubiquitous and tough to eradicate, I strongly feel every child deserves an opportunity to live his/her life to their fullest potential. Countries need to ensure that their children are living in humane conditions and provide the best possible services to them.

Unicef, short for United Nations Children's Fund, is a special program of the United Nations (UN) devoted to aiding national efforts to improve the health, nutrition, education, and general welfare of children. Each year, UNICEF’s flagship publication, The State of the World's Children, closely examines a key issue affecting children. Several statistics for categories such as Basic Indicators, Nutrition, Health, HIV/AIDS, Education, Demographic Indicators, etc. have been provided.

In the first part, I have selected the Education table to find trends and causal factors of the variables. To find the trends I have extensively employed Tableau, which provided me informative visualizations of the data. For the causal factors, I used the Gretl stat tool, which aided me in running quick regressions.

In the second part, I have explored five different hypotheses for different indicators in the tables. More focus was given to improving children's lives in various countries.

In the final part, I have provided relevant recommendations to improve the lives of children. I feel these recommendations will help build a healthier and promising feature for the children.

Indicators

  • Basic Indicators 
  • Nutrition 
  • Health 
  • HIV/AIDS 
  • Education 
  • Demographic Indicators 
  • Economic Indicators 
  • Women 
  • Child Protection 
  • The Rate of Progress 
  • Adolescents 
  • Disparities by Residence 
  • Disparities by household wealth 
  • Early Childhood Development 
  • Under-Five Mortality Rankings

The above data is available on the UNICEF website in the form of a pdf file. I have scraped the data and converted into an excel sheet for data analysis:

As you can see there are over 100 countries in the data set and several columns for each indicator. The data is highly messy with several missing values and alphanumeric characters. We need to come up with legible column names and remove irrelevant columns as well.

Data Cleaning

In order to carry out the above tasks, I have created an effective data cleaning pipeline using python in Google Colab notebook: https://drive.google.com/file/d/1tQS34AaGbb-2M1v6WKlkt9tCVVnrNLZg/view?usp=sharing

Trends

After this process, I decided to take a look at the trends in the Education table as it is one of the most important indicators for improving a child's life. I employed the Gretl Stat tool for finding out the causal factors for each variable in the table and used Tableau for visualizing it.

 

In order to run the regressions in Gretl Stat Tool, I have categorized each country into its specific region and one hot encoded it. Below are a few interesting factors I was able to observe

  1. The youth literacy rate of males and females strongly depends on each other.
  2. The number of internet users per 100 population strongly depends on the lower school enrollment of males and females.
  3. The survival rate of males through primary school is caused mainly by the survival rate of females through primary school.
  4. The mobile phones per 100 population did not seem to have an effect on any of the variables.
  5. The pre-primary school participation decreases significantly when the country is from the South Asia region.

For the full list for causal factors, check here.

Tableau Visualizations

Gross Enrollment Ratio Female

In this graphic, I plotted the gross enrollment ratio for girls in different countries across the world. I found that ‘developed’ countries like the United States, Canada, the UK, and Australia had higher gross enrollments in education for girls compared to ‘developing/under-developed’ nations like South Sudan, the Central African Republic, and Niger. This stands to reason because developed nations are more progressive when it comes to women’s rights and education. Also, in such developed nations, because the level of infrastructure is already higher than that of their less-developed counterparts, it is more feasible for every child to be enrolled in school. 

 

Mobile Phones Per 100 Population 

In this graphic, I plotted the number of mobile phones per 100 population in different countries across the world. I found that countries like Canada, France, the United Kingdom, and Iceland have more mobile phones per 100 as compared to nations like the Democratic Republic of the Congo, Chad, and South Sudan. This makes sense because, in developed nations, people have the money to purchase mobile phones for themselves and other members of their families. However, in less developed nations, people need to prioritize their expenditure because they have lesser disposable income and thus they need to invest all their money in basic necessities. 

 

Internet Users Per 100 Population 

In this graphic, I plotted the number of Internet Users Per 100 Population across the world. I found that countries like Norway, Finland, and Germany have a very large number of Internet users per 100 population as compared to nations like Indonesia, Myanmar, and India. In the more developed European nations, since the quality of infrastructure is better, it is more possible for there to be better Internet connectivity through both WiFi and cellular because in these nations there is more space to build telecom towers and people have the disposable income necessary to purchase WiFi routers or hotspots as compared to poorer nations where the infrastructure is still developing and people don’t have enough disposable income.  

 

Average Youth Literacy

In this graphic, I plotted the average youth literacy for both males and females for all nations. I found that countries like Qatar, Jordan, and Cyprus have a higher percentage of educated youths (male and female) as compared to nations like Uganda, Malawi, and Togo. This makes sense because richer nations have the wherewithal to educate their children while poorer nations have to grapple with providing their children will basic necessities like food and water before they can tackle the education issue

 

Pre-Primary School Participation  

In this graphic, I plotted the Pre-Primary School Participation for both males and females in different nations across the world. I found that nations like Ireland, Portugal, and Denmark have more pre-primary school participation from children as compared to nations like Serbia, Zimbabwe, and Tunisia. This makes sense because most West-European nations have more awareness regarding the importance of education as compared to their less-developed counterparts. 

 

Survival Through Primary School

In this graphic, I plotted the survival through primary school both males and females in different nations across the world. I found that nations like Germany, Japan, and Austria have a higher percentage of children surviving primary school and possibly moving onto the next level of education as compared to nations like Cambodia, Burundi, and Rwanda. This makes sense because of the fact that developed nations tend to emphasize on the importance of education because they are able to due to better quality infrastructure and more disposable income whereas less developed nations cannot do that since they tend to be poorer and the citizens tend to have less disposable income.

Hypothesis Generation

In this section, I have come up with five different hypotheses and performed relevant tests to study their significance.

Hypothesis 1

H0: Countries with lower drinking water service usage have the same percentage of lower secondary school enrollment as compared to their peers.

H1: Countries with lower drinking water services have less enrollment for lower secondary school enrollment as compared to their peers.

Anova One-Way Test

  1. Grouping countries by their water service level:

Result:

The resulting p-value is less than 0.05 and the f-value is around 87. We can reject the null hypothesis and conclude that there is a significant difference between the lower secondary school enrollments for countries with different levels of water service levels. Countries with the lowest drinking water services have a lesser mean of lower secondary school enrollment than their peers which confirms the hypothesis that countries with lower drinking water services have less enrollment for lower secondary school enrollment compared to their peers.

Hypothesis 2

H0: If more children get married at 15 the difference between female and male literacy% remains the same as compared to them marrying between 15-18

H1: If more children get married at 15 the difference between female and male literacy% increases as compared to them marrying between 15-18.

Regression Tests



Result:

The above regressions show us slope is -0.5145 when children marry by 15 and slope is -0.428 when children marry between 15-18. Even though it is negative in both cases, the decrease in the difference between female and male literacy% is lower when children marry between 15-18.

A clearer interpretation would be: for every 1 unit increase in the rate of children marrying by 15, the difference between female and male literacy% decreases by 0.5145 units whereas, for every 1 unit increase in the rate of children marrying between 15-18, the difference between female and male literacy% decreases only by 0.428 units.

Since the p-value is less than 0.05 in both cases, we can reject the null hypothesis at a 95% confidence interval and conclude that if more children get married at 15 the difference between female and male literacy% increases more as compared to them marrying between 15-18.

Hypothesis 3

H0: Countries with Higher Adult Literacy Rate have the same life expectancy as other countries

H1: Countries with Higher Adult Literacy Rate have a higher Life Expectancy than other countries

Regression test: Adult Literacy rate vs Life Expectancy 

Result:

With normalization:

Without normalization:

The above regression shows that ‘Adult Literacy Rate’ has a positive impact on Life Expectancy. And since the p-value of the ‘Adult Literacy Rate’ coefficient is less than 0.05 we can reject the null hypothesis at 95% confidence interval and conclude that Countries with a higher Adult Literacy Rate have a higher Life expectancy.

Hypothesis 4

H0: The Basic Sanitation Usage in urban and rural areas is the same 

H1: The Basic Sanitation Usage in urban and rural areas is not the same 

T-test

Result:

The p-value is less than 0.05 hence we can conclude that the means of Basic Sanitation Usage in Urban and Rural areas are significantly different.

Hypothesis 5

H0: The immunization coverage for all diseases are the same

H1: The immunization coverage for all diseases are not the same

Trends of different immunization coverages:


 ANOVA Test

Result:

We can see the p-value is less than 0.05 and the f-value is around 29.6. Since the p-value is significant we can reject the null hypothesis at the 95% confidence interval and conclude that there the immunization coverage for all diseases is not the same.

The percentage of children receiving ROTA, MCV2 and PCV3 immunization is lower than the percentage of children receiving BCG, Hib3 and DTPI immunization.

Recommendations for improving the quality of children's lives

Females need to marry at a later age: From the hypothesis test between child marriage and literacy rate of youths, we saw that the difference between female and male literacy rate tend to decrease on a smaller scale when females marry between the age of 15-18 than when compared to marriage by the age of 15. Females marrying at a later age will give them a few more years to study which eventually increases their literacy rate and brings them a little closer to the literacy rate of males. Note that this recommendation is based on the context of the data given and does not imply that females only need to marry between the age of 15-18.

Increase the adult literacy rate: By studying the relationship between adult literacy rate and life expectancy at birth we find that countries with more educated adults have a higher life expectancy at birth. Thus, efforts toward adult education programs should be made specially in nations with low literacy rates.

Basic Sanitation Usage in Rural areas should be in line with Urban areasConducting a t-test between Basic Sanitation Usage in Urban areas and Rural areas show that there exists a significant difference between them. This can be in lieu of other factors like lower literacy rates, disparities in basic amenities like electricity, clean drinking water between Rural and Urban areas. This puts children in Rural areas at a disadvantage and higher risk of diseases. 

Countries should increase their ROTA, MCV2 and PCV3 coverage: The relationship between the immunization coverage for various diseases shows that the percentage of children receiving ROTA, MCV2 and PCV3 immunization is lower than the percentage of children receiving BCG, Hib3 and DTPI immunization. Countries should give equal coverage of all immunization types. This will help the health and lifespan of the children.

Countries in Eastern and Southern Africa, and South Asia should better their Internet facilitiesThe trends in the education table show us that if a country is from Eastern and Southern Africa, or South Asia the number of Internet users_per 100 population decreases by 0.33 and 0.42 units respectively. Countries in these regions should expand their internet facilities and encourage internet usage among their population. This will give access to a lot of online resources for the children and help them stay updated on world issues.

Conclusions

After closely examining the UNICEF data about The State of the World's Children, through different analytical tools and techniques I reached the following conclusions:

 

Comments