project_2

War or food?

Are nations dealing with war more food insecure?
How do state expenditures differ on war versus food?

Load in third dataset

File 1 is a file with death rates out of 100k from war link to source
File 2 is a measurement of daily calories consumption link to source
File 3 is GDP War expenditure % link to source
File 4 is GDP Agricultural Expenditure % link to source

delete.na <- function(DF, n=0) {
  DF[rowSums(is.na(DF)) <= n,]
}


War<- read.xlsx("War.xlsx", header=TRUE,1,)
food_security <- read.xlsx("indicator food_consumption.xlsx", header=TRUE,1,)
Agricultural_expend<- read.xlsx("Agricultural_expenditure.xlsx", header=TRUE,1,)
military_expend <- read.xlsx("military_expenditure.xlsx", header=TRUE,1,)

dim(food_security)

## [1] 259  48

dim(military_expend)

## [1] 270  25

dim(Agricultural_expend)

## [1] 270  52

dim(War)

## [1] 192   4

DF Preparation

Eliminate rows with substantial NA’s
Convert columns to numeric
Rename columns to avoid duplication of Column names for eventual Join
Filter df’s for common shared year range of 1988-2007

military_expend <- delete.na(military_expend, 20)
Agricultural_expend <- delete.na(Agricultural_expend, 15)
food_security <- delete.na(food_security, 20)
War <- as_data_frame(lapply(War, function(x) if(is.numeric(x)) round(x,2) else x))
military_expend <- as_data_frame(lapply(military_expend, function(x) if(is.numeric(x)) round(x,2) else x))
War <- War[2:4]
colnames(War) <- c("Country", 'war_2002','war_2004')
colnames(military_expend) <-  c("Country",paste(1988:2011,'Military_expend'))
colnames(food_security) <- c("Country",paste (1961:2007,"food_security"))
colnames(Agricultural_expend) <- c("Country",paste(1961:2011,"Agricultural_expend"))
Agricultural_expend <- Agricultural_expend[c(1,29:48)]
food_security <- food_security[c(1,29:48)]
military_expend <- military_expend[c(1:21)]


merged_df_3<- inner_join(War,military_expend) %>%
    inner_join(.,food_security) %>% 
    inner_join(.,Agricultural_expend)

## Joining, by = "Country"

## Warning: Column `Country` joining factors with different levels, coercing
## to character vector

## Joining, by = "Country"

## Warning: Column `Country` joining character vector and factor, coercing
## into character vector

## Joining, by = "Country"

## Warning: Column `Country` joining character vector and factor, coercing
## into character vector

dim(merged_df_3)

## [1] 83 63

Exploratory Analysis of Death from wars data

Explore war death df

merged_df_3$war_2002-merged_df_3$war_2004

##  [1]  14.90   0.00   0.00   0.00   0.00   0.00   0.00   0.00   0.00   0.00
## [11]   0.00 154.58   0.00   0.01  -0.20   0.00   0.00   0.34  52.60   0.00
## [21]   8.06  88.66  -0.02   0.01   0.00   0.00   0.00   0.00   0.00   0.00
## [31]   0.00   0.00   0.00   0.00   0.02   2.71   0.12   0.00  -0.02   0.00
## [41]   0.00   4.02   1.41   0.00   0.00   0.00   0.00   0.00   0.00   0.00
## [51]   0.00   0.00   0.18 -19.12  -0.02   0.00   0.01   0.00   0.32   0.00
## [61]   0.03   1.55   0.00   0.88  -0.75   9.93   0.00  -0.44  -0.07 -44.86
## [71]   0.00   0.00   0.00   0.00   0.00   0.16   7.46   0.00  -0.07  -0.33
## [81]   0.00   0.00  33.29

merged_df_3[1:3] %>% 
    mutate(.,average_deaths=(war_2002+war_2004)/2) %>% 
    arrange(.,average_deaths) %>%  
    kable(.)

Country	war_2002	war_2004	average_deaths
Argentina	0.00	0.00	0.000
Australia	0.00	0.00	0.000
Austria	0.00	0.00	0.000
Belgium	0.00	0.00	0.000
Benin	0.00	0.00	0.000
Bolivia	0.00	0.00	0.000
Botswana	0.00	0.00	0.000
Brazil	0.00	0.00	0.000
Brunei	0.00	0.00	0.000
Burkina Faso	0.00	0.00	0.000
Cameroon	0.00	0.00	0.000
Chile	0.00	0.00	0.000
China	0.00	0.00	0.000
Cuba	0.00	0.00	0.000
Fiji	0.00	0.00	0.000
Finland	0.00	0.00	0.000
France	0.00	0.00	0.000
Gabon	0.00	0.00	0.000
Gambia	0.00	0.00	0.000
Germany	0.00	0.00	0.000
Ghana	0.00	0.00	0.000
Guyana	0.00	0.00	0.000
Honduras	0.00	0.00	0.000
Ireland	0.00	0.00	0.000
Japan	0.00	0.00	0.000
Lesotho	0.00	0.00	0.000
Madagascar	0.00	0.00	0.000
Malaysia	0.00	0.00	0.000
Mali	0.00	0.00	0.000
Malta	0.00	0.00	0.000
Mauritius	0.00	0.00	0.000
Mexico	0.00	0.00	0.000
New Zealand	0.00	0.00	0.000
Norway	0.00	0.00	0.000
Paraguay	0.00	0.00	0.000
Portugal	0.00	0.00	0.000
Swaziland	0.00	0.00	0.000
Sweden	0.00	0.00	0.000
Togo	0.00	0.00	0.000
Tunisia	0.00	0.00	0.000
United Arab Emirates	0.00	0.00	0.000
Venezuela	0.00	0.00	0.000
Canada	0.01	0.00	0.005
Egypt	0.01	0.00	0.005
Denmark	0.00	0.02	0.010
Hungary	0.01	0.01	0.010
Italy	0.00	0.02	0.010
Netherlands	0.00	0.02	0.010
South Africa	0.01	0.01	0.010
Peru	0.03	0.00	0.015
United Kingdom	0.00	0.07	0.035
Zambia	0.04	0.04	0.040
Niger	0.06	0.05	0.055
Iran	0.12	0.00	0.060
Jordan	0.07	0.07	0.070
Malawi	0.07	0.07	0.070
Turkey	0.16	0.00	0.080
Mauritania	0.09	0.09	0.090
United States	0.01	0.34	0.175
Spain	0.00	0.44	0.220
Saudi Arabia	0.03	0.78	0.405
India	0.53	0.51	0.520
Kuwait	1.41	0.00	0.705
Thailand	0.99	0.99	0.990
Chad	2.03	2.23	2.130
Indonesia	3.98	1.27	2.625
Philippines	4.10	2.55	3.325
Kenya	5.47	1.45	3.460
Pakistan	4.30	3.98	4.140
Sri Lanka	4.88	4.95	4.915
Sierra Leone	9.93	0.00	4.965
Rwanda	9.04	8.16	8.600
Myanmar	9.52	9.34	9.430
Algeria	18.31	3.41	10.860
Nepal	6.25	25.37	15.810
Zimbabwe	33.80	0.51	17.155
Colombia	20.33	19.99	20.160
Congo, Rep.	52.60	0.00	26.300
Cote d’Ivoire	32.25	24.19	28.220
Uganda	43.92	36.46	40.190
Sudan	57.91	102.77	80.340
Congo, Dem. Rep.	130.20	41.54	85.870
Burundi	193.64	39.06	116.350

Tidy dataset

Add average deaths from war column
Create a categorical factor level of deaths from war
- 0 No_Deaths
- 0-1 Low_Deaths
- 1-10 Moderate_Deaths
- 10+ High_Deaths

food_war <- merged_df_3 %>% 
    gather(Wars,Values,4:63) %>% 
    separate(.,Wars,into= c("Year", "Type"),sep=" ") %>% 
    spread(Type,Values) %>% 
    mutate(.,average_deaths=(war_2002+war_2004)/2) %>% 
    select(-c(war_2002,war_2004)) %>% 
    mutate(Deaths_from_war=cut(average_deaths, breaks=c(-1,0,1,10,Inf), labels= c("No_Deaths", "Low_Deaths", "Moderate_Deaths", "High_Deaths")))
   
    

kable(food_war[1:50,])

Country	Year	Agricultural_expend	food_security	Military_expend	average_deaths	Deaths_from_war
Algeria	1988	12.167656	2795.00	1.75	10.86	High_Deaths
Algeria	1989	13.039015	2862.84	1.54	10.86	High_Deaths
Algeria	1990	11.358267	2856.21	1.46	10.86	High_Deaths
Algeria	1991	10.167058	2835.47	1.23	10.86	High_Deaths
Algeria	1992	12.126846	2972.77	2.19	10.86	High_Deaths
Algeria	1993	12.097073	2971.61	2.56	10.86	High_Deaths
Algeria	1994	10.058437	2865.20	3.14	10.86	High_Deaths
Algeria	1995	10.497818	2891.30	2.95	10.86	High_Deaths
Algeria	1996	11.766699	2893.72	3.09	10.86	High_Deaths
Algeria	1997	9.482320	2844.07	3.63	10.86	High_Deaths
Algeria	1998	12.533281	2904.07	3.96	10.86	High_Deaths
Algeria	1999	12.201079	2957.91	3.77	10.86	High_Deaths
Algeria	2000	8.879884	2928.84	3.44	10.86	High_Deaths
Algeria	2001	10.407560	3003.63	3.80	10.86	High_Deaths
Algeria	2002	10.003598	3034.33	3.67	10.86	High_Deaths
Algeria	2003	10.491237	3073.26	3.25	10.86	High_Deaths
Algeria	2004	10.188493	3090.13	3.30	10.86	High_Deaths
Algeria	2005	8.221657	3059.24	2.85	10.86	High_Deaths
Algeria	2006	7.988789	3101.20	2.64	10.86	High_Deaths
Algeria	2007	8.025346	3153.38	2.90	10.86	High_Deaths
Argentina	1988	8.977868	3028.90	1.98	0.00	No_Deaths
Argentina	1989	9.616065	2980.35	1.78	0.00	No_Deaths
Argentina	1990	8.123676	2924.80	1.37	0.00	No_Deaths
Argentina	1991	6.716492	3024.96	1.42	0.00	No_Deaths
Argentina	1992	5.990787	3087.95	1.34	0.00	No_Deaths
Argentina	1993	5.492025	3123.27	1.34	0.00	No_Deaths
Argentina	1994	5.440196	3162.52	1.46	0.00	No_Deaths
Argentina	1995	5.698112	3168.17	1.47	0.00	No_Deaths
Argentina	1996	5.997435	3162.69	1.24	0.00	No_Deaths
Argentina	1997	5.599945	3135.69	1.14	0.00	No_Deaths
Argentina	1998	5.617612	3166.62	1.14	0.00	No_Deaths
Argentina	1999	4.735356	3266.75	1.22	0.00	No_Deaths
Argentina	2000	4.971221	3271.70	1.15	0.00	No_Deaths
Argentina	2001	4.814450	3183.34	1.18	0.00	No_Deaths
Argentina	2002	10.683021	2965.90	1.09	0.00	No_Deaths
Argentina	2003	10.986832	3059.74	1.06	0.00	No_Deaths
Argentina	2004	10.412916	3029.91	0.96	0.00	No_Deaths
Argentina	2005	9.401029	3091.79	0.93	0.00	No_Deaths
Argentina	2006	8.392957	2969.12	0.86	0.00	No_Deaths
Argentina	2007	9.394875	2940.98	0.88	0.00	No_Deaths
Australia	1988	5.208421	3121.54	2.28	0.00	No_Deaths
Australia	1989	5.364690	3141.58	2.15	0.00	No_Deaths
Australia	1990	4.923955	3177.79	2.10	0.00	No_Deaths
Australia	1991	3.631461	3109.40	2.15	0.00	No_Deaths
Australia	1992	3.520368	3116.99	2.19	0.00	No_Deaths
Australia	1993	3.710071	3044.14	2.21	0.00	No_Deaths
Australia	1994	3.799483	3038.66	2.17	0.00	No_Deaths
Australia	1995	3.416733	3081.55	2.07	0.00	No_Deaths
Australia	1996	3.819477	3056.62	1.97	0.00	No_Deaths
Australia	1997	3.666767	3091.38	1.91	0.00	No_Deaths

Summary Explorations

food_war <-food_war %>% 
    dplyr::group_by(Deaths_from_war,Year)
agricultural_spending_by_year <- food_war %>% 
    dplyr::summarize(agri_exp_mean=mean(Agricultural_expend, na.rm = TRUE)) 
food_war %>% 
    dplyr::summarize(Military_Spending_Pct_GDP_By_Year=mean(Military_expend, na.rm = TRUE)) %>% 
    kable(.)

Deaths_from_war	Year	Military_Spending_Pct_GDP_By_Year
No_Deaths	1988	2.323429
No_Deaths	1989	2.268684
No_Deaths	1990	2.363947
No_Deaths	1991	2.218108
No_Deaths	1992	1.989167
No_Deaths	1993	1.972105
No_Deaths	1994	1.914865
No_Deaths	1995	1.834324
No_Deaths	1996	1.737778
No_Deaths	1997	1.873889
No_Deaths	1998	1.905556
No_Deaths	1999	1.811622
No_Deaths	2000	1.734000
No_Deaths	2001	1.732250
No_Deaths	2002	1.673750
No_Deaths	2003	1.719286
No_Deaths	2004	1.628095
No_Deaths	2005	1.590476
No_Deaths	2006	1.501951
No_Deaths	2007	1.528750
Low_Deaths	1988	4.386000
Low_Deaths	1989	4.038095
Low_Deaths	1990	5.793333
Low_Deaths	1991	8.931429
Low_Deaths	1992	4.616191
Low_Deaths	1993	3.623810
Low_Deaths	1994	3.422273
Low_Deaths	1995	3.093182
Low_Deaths	1996	2.920000
Low_Deaths	1997	2.936818
Low_Deaths	1998	3.098095
Low_Deaths	1999	2.916364
Low_Deaths	2000	2.933810
Low_Deaths	2001	3.063810
Low_Deaths	2002	2.793810
Low_Deaths	2003	2.787619
Low_Deaths	2004	2.625455
Low_Deaths	2005	2.495000
Low_Deaths	2006	2.490476
Low_Deaths	2007	2.481500
Moderate_Deaths	1988	2.511250
Moderate_Deaths	1989	2.498750
Moderate_Deaths	1990	2.955000
Moderate_Deaths	1991	3.241250
Moderate_Deaths	1992	3.203750
Moderate_Deaths	1993	3.255556
Moderate_Deaths	1994	2.980000
Moderate_Deaths	1995	3.244444
Moderate_Deaths	1996	3.170000
Moderate_Deaths	1997	2.593333
Moderate_Deaths	1998	2.661250
Moderate_Deaths	1999	2.536250
Moderate_Deaths	2000	2.921250
Moderate_Deaths	2001	2.497778
Moderate_Deaths	2002	2.276667
Moderate_Deaths	2003	2.266250
Moderate_Deaths	2004	2.028750
Moderate_Deaths	2005	1.893750
Moderate_Deaths	2006	2.001429
Moderate_Deaths	2007	2.451250
High_Deaths	1988	2.592500
High_Deaths	1989	2.520000
High_Deaths	1990	2.518750
High_Deaths	1991	2.283750
High_Deaths	1992	2.321250
High_Deaths	1993	2.395000
High_Deaths	1994	2.254286
High_Deaths	1995	2.436667
High_Deaths	1996	2.593333
High_Deaths	1997	2.473333
High_Deaths	1998	2.882857
High_Deaths	1999	3.325000
High_Deaths	2000	3.235000
High_Deaths	2001	3.190000
High_Deaths	2002	4.026250
High_Deaths	2003	2.690000
High_Deaths	2004	3.136000
High_Deaths	2005	2.664000
High_Deaths	2006	2.562000
High_Deaths	2007	2.337778

food_security_by_year <- food_war %>% 
    dplyr::summarize(my_mean=mean(food_security, na.rm = TRUE))

food_war <-food_war %>% 
    dplyr::group_by(Deaths_from_war)
food_war %>% 
    dplyr::summarize(Mean_Agricultural_Spending_Pct_GDP=mean(Agricultural_expend, na.rm = TRUE)) %>% 
    kable(.)

Deaths_from_war	Mean_Agricultural_Spending_Pct_GDP
No_Deaths	13.00998
Low_Deaths	11.75719
Moderate_Deaths	31.46913
High_Deaths	29.13588

food_war %>% 
    dplyr::summarize(Mean_Military_Spending_Pct=mean(Military_expend, na.rm = TRUE)) %>% 
    kable(.)

Deaths_from_war	Mean_Military_Spending_Pct
No_Deaths	1.856862
Low_Deaths	3.562188
Moderate_Deaths	2.671807
High_Deaths	2.725783

food_war %>% 
    dplyr::summarize(mean_Caloric_consuimption=mean(food_security, na.rm = TRUE)) %>% 
    kable(.)

Deaths_from_war	mean_Caloric_consuimption
No_Deaths	2827.043
Low_Deaths	2913.194
Moderate_Deaths	2146.651
High_Deaths	2234.371

Warnings about observations

My categories aren’t of even sizes so my clustering could effect my observations and lead to large impact from outliers
I don’t have an overall GDP per country, this could make a big difference in relative % GDP expenditures
Tentative Observations
- Areas with the most death from wars seem to spend more of their GDP on agricultural expenditures
- Areas with the most deaths don’t spend much on military expenditures, and low death areas actually spend more than high death areas
- High and moderate death areas are significantly more food insecure. Calorie consumption is about 30% lower in these areas

Data Visualization

food_security_by_year

## # A tibble: 80 x 3
## # Groups: Deaths_from_war [?]
##    Deaths_from_war Year  my_mean
##    <fctr>          <chr>   <dbl>
##  1 No_Deaths       1988     2744
##  2 No_Deaths       1989     2741
##  3 No_Deaths       1990     2734
##  4 No_Deaths       1991     2745
##  5 No_Deaths       1992     2766
##  6 No_Deaths       1993     2758
##  7 No_Deaths       1994     2772
##  8 No_Deaths       1995     2780
##  9 No_Deaths       1996     2801
## 10 No_Deaths       1997     2809
## # ... with 70 more rows

food_security_by_year %>% 
    ggplot(.,aes(x=Year,y=my_mean))+
    geom_bar(aes(fill = Deaths_from_war), position = "dodge", stat = "identity")+
    theme(axis.text.x=element_text(angle=45,hjust=1))

Caloric consumption seems rather Flat in all the categories over the last three decades

agricultural_spending_by_year %>% 
    ggplot(.,aes(x=Year,y=agri_exp_mean))+
    geom_bar(aes(fill = Deaths_from_war), position = "dodge", stat = "identity")+
    theme(axis.text.x=element_text(angle=45,hjust=1))

As a percent of GDP agricultural expenditure has decreased across all categories over last three decades
Area’s with higher casualties of war seem to have made large cuts to expenditures

Most interesting graphs and observations

A handful of outliers were removed for better visualization

food_war %>% 
    ggplot(.,aes(x=Agricultural_expend,y=food_security,shape=Deaths_from_war,color=Deaths_from_war))+
    geom_point()

## Warning: Removed 15 rows containing missing values (geom_point).

Graph 1 Agricultural Exp vs Food Security

Two Clear Clusters develop in the first graph
- Areas with high death rates seem to be the least food secure and simultaneously spend the most on agricultural subsidy
- Areas with low death rates spends the least on food subsidy, and are the most food secure

food_war %>% 
    filter(.,Military_expend <20) %>% 
    ggplot(.,aes(x=Military_expend,y=food_security,shape=Deaths_from_war,color=Deaths_from_war))+
    geom_point()

Graph 2 Military Exp vs Food Security

Areas with low death rates seem to spend the most on the Military

project_2_c

Justin Herman

March 11, 2018

Third Dataset- Gapminder

War or food?

Load in third dataset

DF Preparation

Exploratory Analysis of Death from wars data

Tidy dataset

Summary Explorations

Warnings about observations

Data Visualization

Most interesting graphs and observations

Graph 1 Agricultural Exp vs Food Security

Graph 2 Military Exp vs Food Security