Life Expectancy

The data set

I am using the gapminder data for this project.

First look at the data

Open the data as a dataframe and look at the shape

import pandas as pd
df = pd.read_csv("../../../../data/gapminder.csv")
s = df.shape 
print(f'There are {s[0]} rows and {s[1]} columns')
c_names = df.columns.tolist()
print("The headings are: ")
for c in c_names:
    print(f'- {c}')

There are 1704 rows and 6 columns
The headings are: 
- country
- year
- pop
- continent
- lifeExp
- gdpPercap

Get a summary of the numerical data

df.describe()

	year	pop	lifeExp	gdpPercap
count	1704.00000	1.704000e+03	1704.000000	1704.000000
mean	1979.50000	2.960121e+07	59.474439	7215.327081
std	17.26533	1.061579e+08	12.917107	9857.454543
min	1952.00000	6.001100e+04	23.599000	241.165876
25%	1965.75000	2.793664e+06	48.198000	1202.060309
50%	1979.50000	7.023596e+06	60.712500	3531.846988
75%	1993.25000	1.958522e+07	70.845500	9325.462346
max	2007.00000	1.318683e+09	82.603000	113523.132900

Check for null values and data types.

df.info()
print("There are no null values that would have been removed in data cleaning.")

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1704 entries, 0 to 1703
Data columns (total 6 columns):
 #   Column     Non-Null Count  Dtype  
---  ------     --------------  -----  
 0   country    1704 non-null   object 
 1   year       1704 non-null   int64  
 2   pop        1704 non-null   float64
 3   continent  1704 non-null   object 
 4   lifeExp    1704 non-null   float64
 5   gdpPercap  1704 non-null   float64
dtypes: float64(3), int64(1), object(2)
memory usage: 80.0+ KB
There are no null values that would have been removed in data cleaning.

Unique values for potential catagories

Check Country first

all_countries = df.country.unique().tolist() 
print(f'This data has {len(all_countries)} different countries')

This data has 142 different countries

Check Continent next

all_continents = df.continent.unique().tolist()
print(f"This data has {len(all_continents)} different continents")

This data has 5 different continents

Also check years

all_years = df.year.unique().tolist()
print(f"This data has {len(all_years)} different years")
for y in all_years:
    print(y)

This data has 12 different years
1952
1957
1962
1967
1972
1977
1982
1987
1992
1997
2002
2007

Summary

We have data from 1952 to 2007 accross 142 countries and 5 continents. Each of these has population, life expectancy and GDP. I will see if I can find any links for life expectancy based on years.

Years vs Life Expectancy

import seaborn as sns
sns.catplot(data=df, x="year", y="lifeExp", kind="box")

It appears that life expectancy increased with time but I will do a deeper dive

Mean life expectancy by years

gb = df.groupby("year")["lifeExp"].mean()
data = {"years" : all_years, "life" : gb }
doggy = pd.DataFrame(data)
sns.relplot(data = doggy, x = "years", y = "life", kind = "line")

Is the difference significant?

Null hypothesis - there is no difference in life expectancy between 1957 and 2007

I will perform a ttest to check this

import scipy.stats as stats
early = df[df.year == 1957]
late = df[df.year == 2007]
check = stats.ttest_ind(early.lifeExp, late.lifeExp)
print(f'The pvalue is {check.pvalue}.')

The pvalue is 8.299952773896099e-23.

Result

There is a very low pvalue indcating that we can reject the null hypothesis there is no difference in life expectancy between 1952 and 2007.

Look at outlier 1992

life_year = df[(df.year == 1992)]
min_exp_92 = life_year.lifeExp.min()
min_country = life_year.country[life_year.lifeExp == min_exp_92].tolist()
print(f'The 1992 outlier is {str(min_country[0])} with a life expectancy of {min_exp_92}')

The 1992 outlier is Rwanda with a life expectancy of 23.599

In 1992 there was a 100 day civil war in Rwanda that led to the genocide of nearly 1 million people. Source