In [42]:
total['GDP'].plot.bar();
This project will find out if a country with the highest GDP per capita would have a higher percentage of refugees. The notion is a country that is wealthier would have the necessary resources and fund to host as many refugees that the one that has the lowest GDP per capita. This analysis will not be a comprehensive one as GDP per capita depends on a number of factors including population size.
For this project, I will be including three indicators; population growth, refugees by percentage* and GDP per capita. I will be focusing on three countries which I will choose based on the lowest, middle and highest GDP per capita in the year 2017. The focus year is 2017 to simplify the research. All data will be extracted from Gapminder.com.
NOTE*: the percentage of refugees in a host country abroad. This data tells us about refugees living outside of their home country.
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sb
%matplotlib inline
refugee_number = pd.read_csv('refugee_in_perc_wb.csv')
refugee_number.set_index('Country', drop = True, inplace= True)
gdp_per_capita = pd.read_csv('income_per_person_gdp.csv')
gdp_per_capita.set_index('Country', drop = True, inplace= True)
population_growth = pd.read_csv('population_growth_annual_percent.csv')
population_growth.set_index('country', drop = True, inplace = True)
population_growth.info(), gdp_per_capita.info(), refugee_number.info()
<class 'pandas.core.frame.DataFrame'> Index: 194 entries, Afghanistan to Zimbabwe Data columns (total 58 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 1960 191 non-null float64 1 1961 192 non-null float64 2 1962 192 non-null float64 3 1963 192 non-null float64 4 1964 192 non-null float64 5 1965 192 non-null float64 6 1966 192 non-null float64 7 1967 192 non-null float64 8 1968 192 non-null float64 9 1969 192 non-null float64 10 1970 192 non-null float64 11 1971 192 non-null float64 12 1972 192 non-null float64 13 1973 192 non-null float64 14 1974 192 non-null float64 15 1975 192 non-null float64 16 1976 192 non-null float64 17 1977 192 non-null float64 18 1978 192 non-null float64 19 1979 192 non-null float64 20 1980 192 non-null float64 21 1981 192 non-null float64 22 1982 192 non-null float64 23 1983 192 non-null float64 24 1984 192 non-null float64 25 1985 192 non-null float64 26 1986 192 non-null float64 27 1987 192 non-null float64 28 1988 192 non-null float64 29 1989 192 non-null float64 30 1990 193 non-null float64 31 1991 193 non-null float64 32 1992 193 non-null float64 33 1993 193 non-null float64 34 1994 193 non-null float64 35 1995 193 non-null float64 36 1996 194 non-null float64 37 1997 194 non-null float64 38 1998 194 non-null float64 39 1999 194 non-null float64 40 2000 194 non-null float64 41 2001 194 non-null float64 42 2002 194 non-null float64 43 2003 194 non-null float64 44 2004 194 non-null float64 45 2005 194 non-null float64 46 2006 194 non-null float64 47 2007 194 non-null float64 48 2008 194 non-null float64 49 2009 194 non-null float64 50 2010 194 non-null float64 51 2011 193 non-null float64 52 2012 193 non-null float64 53 2013 193 non-null float64 54 2014 193 non-null float64 55 2015 193 non-null float64 56 2016 193 non-null float64 57 2017 193 non-null float64 dtypes: float64(58) memory usage: 88.7+ KB <class 'pandas.core.frame.DataFrame'> Index: 193 entries, Afghanistan to Zimbabwe Columns: 241 entries, 1800 to 2040 dtypes: int64(241) memory usage: 364.1+ KB <class 'pandas.core.frame.DataFrame'> Index: 194 entries, Afghanistan to Zimbabwe Data columns (total 29 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 1990 194 non-null float64 1 1991 194 non-null float64 2 1992 194 non-null float64 3 1993 194 non-null float64 4 1994 194 non-null float64 5 1995 194 non-null float64 6 1996 194 non-null float64 7 1997 194 non-null float64 8 1998 194 non-null float64 9 1999 194 non-null float64 10 2000 194 non-null float64 11 2001 194 non-null float64 12 2002 194 non-null float64 13 2003 194 non-null float64 14 2004 194 non-null float64 15 2005 194 non-null float64 16 2006 194 non-null float64 17 2007 194 non-null float64 18 2008 194 non-null float64 19 2009 194 non-null float64 20 2010 194 non-null float64 21 2011 194 non-null float64 22 2012 194 non-null float64 23 2013 194 non-null float64 24 2014 194 non-null float64 25 2015 194 non-null float64 26 2016 194 non-null float64 27 2017 194 non-null float64 28 2018 194 non-null float64 dtypes: float64(29) memory usage: 44.7+ KB
(None, None, None)
There are some missing data and they will be fixed in the cleaning process next.
refugee_number.drop(refugee_number.loc[:,:'2016'].columns, axis = 1, inplace= True) #to drop all years except 2017
refugee_number.drop(['2018'], axis = 1, inplace = True)
gdp_per_capita.drop(gdp_per_capita.loc[:,:'2016'].columns, axis = 1, inplace= True)
gdp_per_capita.drop(gdp_per_capita.loc[:, '2018':].columns, axis = 1, inplace= True)
population_growth.drop(population_growth.loc[:,:'2016'].columns, axis = 1, inplace= True)
gdp_per_capita.head()
| 2017 | |
|---|---|
| Country | |
| Afghanistan | 1760 |
| Albania | 11800 |
| Algeria | 13900 |
| Andorra | 49800 |
| Angola | 6040 |
As mentioned in the introduction, the analysis will be on the data in the year 2017. This is just to simplify the project. I have removed all other years using the slicing method. As seen the output, only data in one year for each indicator will be analysed.
(refugee_number.duplicated().sum().sum(), gdp_per_capita.duplicated().sum().sum(), population_growth.duplicated().sum().sum())
(38, 16, 27)
There are duplicates. However, after a close examination of the Excel sheet, I found the 'duplicated' data are not duplicated. Most were rounded off to 2 decimals but the other to 3 decimals. When I kept the numbers constant to 3 decimals, the 'duplicated' data did not alter. In any case, the duplicated data will not be removed.
(refugee_number.isna().sum().sum(), gdp_per_capita.isna().sum().sum(), population_growth.isna().sum().sum())
(0, 0, 1)
population_growth.dropna(inplace= True)
population_growth.isna().sum().sum()
0
All NaN removed. The next steps will be to decide the three countries for analysis. They will be chosen based on their GDP per capita.
gdp_per_capita.describe()
| 2017 | |
|---|---|
| count | 193.000000 |
| mean | 18172.699482 |
| std | 19393.559289 |
| min | 627.000000 |
| 25% | 3650.000000 |
| 50% | 11700.000000 |
| 75% | 26400.000000 |
| max | 113000.000000 |
# To determine the country with the lowest, highest and median GDP per capita:
gdp_countries = gdp_per_capita[gdp_per_capita['2017'].isin(["627", "113000", "11700"])]
gdp_countries
# So the three countries are as follow:
| 2017 | |
|---|---|
| Country | |
| Qatar | 113000 |
| Somalia | 627 |
| Sri Lanka | 11700 |
refugee = refugee_number.loc[['Somalia', 'Qatar', 'Sri Lanka']]
refugee.rename(columns = {'2017': 'Refugee'}, inplace=True)
population = population_growth.loc[['Somalia', 'Qatar', 'Sri Lanka']]
population.rename(columns = {'2017': 'Population'}, inplace = True)
gdp = gdp_countries.loc[['Somalia', 'Qatar', 'Sri Lanka']]
gdp.rename(columns = {'2017': 'GDP'}, inplace = True)
mean_gdp = gdp_per_capita.mean()
meanGDP = mean_gdp.to_frame().T
I used the location method to find the three countries in each indicator i.e. refugee percentage, population growth and GDP per capita. I have then renamed all the indicated above. I had an initial idea of using the mean GDP per capita as a gauge (dependent). I have managed to create a separate dataframe for the mean but I got into a pickle while plotting the graphs.
total = gdp.join([refugee, population])
total
| GDP | Refugee | Population | |
|---|---|---|---|
| Country | |||
| Somalia | 627 | 0.09980 | 2.92 |
| Qatar | 113000 | 0.00690 | 2.67 |
| Sri Lanka | 11700 | 0.00388 | 1.13 |
The idea is to amalgamate the indicators as one table to simplify plotting later.
total['GDP'].plot.bar();
Based on the visual above, it tells me that Qatar's GDP per capita is signifcantly higher than the other two countries. Sri Lanka and Somalia have GDP per capita, collectively, under 20000, while Qatar's shoot overs.
total['Population'].plot.bar();
Somalia has a higher population growth than Qatar. Sri Lanka has the least. But the growth between Qatar and Somalia are not far from each other.
total['Refugee'].plot.bar();
I was quite surprised by this graph result. Somalia has the least GDP per capita but its percentage of refugees is the highest as compared to Qatar's, which has the highest amongst countries. Sri Lanka has the least share. A least wealthy country has hosted the most refugees.
total.plot.bar(y = ['Refugee', 'Population']);
I have plotted this to see side by side, the refugee percentage and population growth, to see the magnitude of the latter when compared to the former.
total.plot.bar(y = ['GDP', 'Population',]);
Because the values for GDP is just too great as compared to the population growth, one cannot see its values on the graph.
total.plot(kind='scatter',x='Population',y='Refugee',color='red')
plt.show()
total.plot(kind='scatter',x='GDP',y='Refugee',color='red')
plt.show()
The country that has the least population is taking in fewer refugees. As confirmed in the bar chart above, Somalia that has the most growth has the most percentage of refugees. I would imagine countries like Qatar and Sri Lanka that the least population growth and higher GDP per capita would consider taking in more refugees.
The countries that have least population growth have fewer refugees. The countries that have lower GDP per capita have hosted the most percentage of refugees. The wealthiest country has fewer refugees. This finding tells that the country that has the highest GDP per capita has the fewest number of refugees.
Limitation: Ideally, the mean GDP should be incorporated to see where these countries stand when compared to the mean.
https://www.listendata.com/2019/07/how-to-filter-pandas-dataframe.html
https://pandas.pydata.org/pandas-docs/stable/reference/frame.html
https://www.tutorialspoint.com/numpy/index.htm
https://pandas.pydata.org/pandas-docs/stable/user_guide/visualization.html
http://queirozf.com/entries/pandas-dataframe-plot-examples-with-matplotlib-pyplot
!jupyter nbconvert "Exploration of Loan Data from Prosper.ipynb" --to html --output-dir='C:\Users\Ajeet\OneDrive\Projects\Blogging\ajeethaa.github.io'
C:\Users\Ajeet\anaconda3\lib\site-packages\traitlets\traitlets.py:2195: FutureWarning: Supporting extra quotes around Unicode is deprecated in traitlets 5.0. Use 'C:\\Users\\Ajeet\\OneDrive\\Projects\\Blogging\\ajeethaa.github.io' instead of "'C:\\Users\\Ajeet\\OneDrive\\Projects\\Blogging\\ajeethaa.github.io'" – or use CUnicode. warn( [NbConvertApp] Converting notebook Exploration of Loan Data from Prosper.ipynb to html [NbConvertApp] Writing 1192320 bytes to C:\Users\Ajeet\OneDrive\Projects\Blogging\ajeethaa.github.io\Exploration of Loan Data from Prosper.html