AgTrendsReport.Rmd

---
title: "Uganda Agricultural Trends Report"
author: "Pardee"
date: "Spring/Summer 2018"
output:
  html_notebook:
    toc: yes
  html_document:
    toc: yes
    toc_depth: 3
---

```{r setup, include=FALSE}
library(tidyverse)
library(scales)
library(extrafont)
library(ggalt)            #for easy dumbbell plots
#library(stringr)
#font_import()
#loadfonts()

pardee.theme <- theme(
  plot.title = element_text(family="Gill Sans MT", color="black", face="bold", size=16, hjust=0), 
  plot.subtitle = element_text(family="Gill Sans MT", color="black", size=14, hjust=0), 
  axis.title = element_text(family="Gill Sans MT", color="black", face="bold", size=12), 
  axis.text = element_text(family="Gill Sans MT", color="black", size=11), 
  plot.caption = element_text(family="Gill Sans MT", color="black", size=12), 
  panel.background =  element_rect(fill = "#F7F7F7", colour = NA), 
  panel.grid.major = element_line(colour = "grey90", size = 0.5),
  panel.grid.minor = element_line(colour = "grey93", size = 0.5),
  panel.border = element_rect(colour = "black", size = 0.5, fill=NA, linetype = 1),
  legend.title=element_blank(), 
  legend.text = element_text(family="Gill Sans MT", color="black", size=12, hjust=0),
  legend.position = 'top',
  strip.text = element_text(family="Gill Sans MT", color="black", size=14))

pardee.theme2 <- theme(
  plot.title = element_text(color="black", face="bold", size=16, hjust=0), 
  plot.subtitle = element_text(color="black", size=14, hjust=0), 
  axis.title = element_text(color="black", face="bold", size=12), 
  axis.text = element_text(color="black", size=11), 
  plot.caption = element_text(color="black", size=12), 
  panel.background =  element_rect(fill = "#F7F7F7", colour = NA), 
  panel.grid.major = element_line(colour = "grey90", size = 0.5),
  panel.grid.minor = element_line(colour = "grey93", size = 0.5),
  panel.border = element_rect(colour = "black", size = 0.5, fill=NA, linetype = 1),
  legend.title=element_blank(), 
  legend.text = element_text(color="black", size=12, hjust=0),
  legend.position = 'top',
  strip.text = element_text(color="black", size=14))

viz_cols <- c('Uganda' = 'red', 'Other' = 'black', 'Rest of World' = 'lightgray',
              'Current Path' = 'darkgreen', 'Food Secure' = 'tomato3', 'Agricultural Production' = 'blue')

pt_groups <- read.csv('.//data/pardee_trends_groups.csv') %>% 
  filter(main_group == 'Pardee Group')
```

# Population growth

## Doubling time 

```{r uganda_pop, include=TRUE}
uganda_pop <- readxl::read_xls('.//data/UGANDA_POP.xls') %>% 
  mutate(pop_times_2 = Population * 2,
         dbl_1960 = ifelse(Population > pop_times_2[1], 1, 0),
         dbl_1960_year = ifelse(dbl_1960 == 1, which.min(Year), NA))

uganda_pop$pop_times_2[1]
```

```{r eac_countries, include=TRUE}
eac_countries <- c('Burundi', 'Kenya', 'Rwanda', 'Tanzania', 'Uganda')
```

# Median age

```{r median_age, include=TRUE}
median_age <- readxl::read_xls('.//data/MEDIAN_AGE_AFRICA.xls') %>% 
  filter(Year == 2018 | Year == 2050) %>% 
  mutate(IsUganda = ifelse(Country == 'Uganda', 'Uganda', 'Other'),
         Context = 'Countries')
```

```{r median_age_eac, include=TRUE}
median_age_eac <- median_age %>% 
  filter(Country %in% eac_countries)
```

```{r median_age_context, include=TRUE}
median_age_context <- readxl::read_xls('.//data/CONTEXT_MEDIAN_AGE.xls') %>% 
  filter(Year == 2018 | Year == 2050) %>% 
  mutate(IsUganda = 'Other',
         Context = 'Regions')

test <- bind_rows(median_age_eac, median_age_context)
```

```{r fig.height=6.2}
ggplot() + 
  geom_point(data=median_age %>% filter(Year == 2018 & 
                                        Country %in% eac_countries),
             aes(x=MedianAge,
                 y=reorder(Country, -MedianAge),
                 color=IsUganda), size=5) + 
  geom_point(data=median_age %>% filter(Year == 2050 & 
                                        Country %in% eac_countries),
             aes(x=MedianAge,
                 y=reorder(Country, -MedianAge),
                 color=IsUganda), size=5) + 
  scale_color_manual(values = viz_cols) + 
  labs(y='') + 
  pardee.theme
```


```{r}
median_age_eac %>% 
  spread(Year, MedianAge) %>% 
  ggplot(.) + 
  geom_dumbbell(aes(x=`2018`,
                    xend=`2050`,
                    y=reorder(Country, -`2018`),
                    color=IsUganda), size_x = 5, size_xend = 5, dot_guide = TRUE) + 
  ggtitle('Median age of EAC countries in 2018 and 2050') + 
  annotate("text", x = 16, y = 5.5, label = "2018", 
           color = 'gray40', size = 4.5, family="Gill Sans MT", fontface = 'bold') + 
  annotate("text", x = 22, y = 5.5, label = "2050", 
           color = 'gray40', size = 4.5, family="Gill Sans MT", fontface = 'bold') + 
  scale_color_manual(values = viz_cols) + 
  labs(x='Median age of population',
       y='') + 
  scale_y_discrete(expand = c(0,1)) +
  pardee.theme 
  theme(legend.position = 'none')
```


```{r fig.height=4.8, fig.width=10}
test %>% 
  filter(!str_detect(Country, 'Africa'),
         Country != 'Pardee LatAm & Car') %>% 
  mutate(viz_facet = ifelse(Context == 'Countries', 'East African Community', 'UN Regions')) %>% 
  filter(viz_facet == 'East African Community') %>% 
  spread(Year, MedianAge) %>% 
  ggplot(.) + 
  geom_dumbbell(aes(x=`2018`,
                    xend=`2050`,
                    y=reorder(Country, -`2018`),
                    color=IsUganda), size_x = 5, size_xend = 5, dot_guide = TRUE) + 
  ggtitle('Median Age of countries and regions in 2018 and 2050',
          subtitle = 'IFs Current Path') +
  annotate("text", x = 16, y = 5.3, label = "2018",  
           color = 'gray40', size = 4.5, family="Gill Sans MT", fontface = 'bold') + 
  annotate("text", x = 22.25, y = 5.3, label = "2050", 
           color = 'gray40', size = 4.5, family="Gill Sans MT", fontface = 'bold') + 
  facet_wrap(~viz_facet, ncol = 1, scales = "free_y") + 
  scale_color_manual(values = viz_cols) + 
  labs(x='',
       y='') + 
  pardee.theme + 
  theme(legend.position = 'none')


```

```{r fig.height=4.8, fig.width=10}
test %>% 
  filter(!str_detect(Country, 'Africa'),
         Country != 'Pardee LatAm & Car') %>% 
  mutate(viz_facet = ifelse(Context == 'Countries', 'East African Community', 'UN Regions')) %>% 
  filter(viz_facet == 'UN Regions') %>% 
  spread(Year, MedianAge) %>% 
  ggplot(.) + 
  geom_dumbbell(aes(x=`2018`,
                    xend=`2050`,
                    y=reorder(Country, -`2018`),
                    color=IsUganda), size_x = 5, size_xend = 5, dot_guide = TRUE) + 
  #ggtitle('Median Age of countries in 2018 and 2050',
   #       subtitle = 'IFs Current Path') +
  #annotate("text", x = 16, y = 5.3, label = "2018",  
   #        color = 'gray40', size = 4.5, family="Gill Sans MT", fontface = 'bold') + 
  #annotate("text", x = 22.25, y = 5.3, label = "2050", 
   #        color = 'gray40', size = 4.5, family="Gill Sans MT", fontface = 'bold') + 
  facet_wrap(~viz_facet, ncol = 1, scales = "free_y") + 
  scale_color_manual(values = viz_cols) + 
  labs(x='Median age of population',
       y='') + 
  pardee.theme + 
  theme(legend.position = 'none')
```

```{r population, include=TRUE}
population <- read_csv('.//data/pop_story.csv') %>% 
  gather(year, val, 5:94)

population <- population %>% 
  filter(year == 2018) %>% 
  select(-unit) %>% 
  spread(var, val) %>% 
  mutate(IsUganda = ifelse(country == 'Uganda', 'Uganda', 'Rest of World'))
```

```{r population_2, include=TRUE}
population <- population %>% 
  left_join(x=.,
            y=pt_groups,
            by=c("country" = "Country"))
```

```{r pop_ttl, include=TRUE}
pop_ttl <- readxl::read_xlsx('.//data/population_ttl.xlsx') %>% 
  gather(year, population_ttl, 5:40) %>% 
  #mutate(year = as.numeric(year)) %>% 
  filter(year == '2018') %>% 
  select(country, scenario, year, population_ttl)

population <- population %>% 
  left_join(x=.,
            y=pop_ttl,
            by='country')

#write.csv(population, './/output_tables/ug_fig_2.csv')
```

```{r fig.height=6.2}
fig_2 <- population %>% 
  ggplot(.) + 
  geom_point(aes(x=pop_med,
                 y=pop_gr / 100,
                 color=sub_group, size=population_ttl)) + 
  ggrepel::geom_text_repel(data=population %>% filter(country %in% eac_countries),
                            aes(x=pop_med,
                                y=pop_gr / 100,
                                label=country), 
                           family="Gill Sans MT", 
                           size=4.5, force = 24) + 
  ggtitle('Median age of population vs. annual population growth',
          'Demographic indicators from 2018') + 
  labs(x='\nMedian age of population',
       y='Annual population growth rate\n',
       caption='Source: IFs v7.33 Current Path\nBubble size indicates the size\nof the total population') + 
  scale_y_continuous(labels = percent,
                     limits = c(-.025, .052)) + 
  scale_x_continuous(limits = c(15, 48)) + 
  pardee.theme + 
  guides(size= 'none') + 
  theme(axis.text = element_text(size=16))

ggsave(".//graphics/fig_2.pdf", plot = fig_2, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

```{r ag_import_dep, include=TRUE}
ag_import_dep <- readxl::read_xlsx('.//data/report_graphics.xlsx', sheet='ag_import_dep') %>% 
  gather(year, val, 5:50) %>% 
  mutate(year = as.numeric(year),
         scenario = factor(scenario, levels=c('Current Path', 'Agricultural Production', 'Food Secure')))

maln_pop <- readxl::read_xlsx('.//data/report_graphics.xlsx', sheet='maln_pop') %>% 
  gather(year, val, 5:64) %>% 
  mutate(year = as.numeric(year),
         scenario = factor(scenario, levels=c('Current Path', 'Agricultural Production', 'Food Secure')))

yyl <- readxl::read_xlsx('.//data/report_graphics.xlsx', sheet='yyl') %>% 
  gather(year, val, 5:40) %>% 
  mutate(year = as.numeric(year),
         scenario = factor(scenario, levels=c('Current Path', 'Agricultural Production', 'Food Secure')))

cum_gdp <- readxl::read_xlsx('.//data/report_graphics.xlsx', sheet='cum_gdp') %>% 
  gather(year, val, 5:37) %>% 
  mutate(year = as.numeric(year),
         scenario = factor(scenario, levels=c('Current Path', 'Agricultural Production', 'Food Secure')))
```

```{r trade_cats, include=TRUE}
quantity_vals <- c('Export Quantity', 'Import Quantity')
value_vals <- c('Export Value', 'Import Value')
```

```{r ag_trade, include=TRUE}
ag_trade <- readxl::read_xlsx('.//data/ag_trade.xlsx') %>% 
  filter(Country == 'Uganda')

ag_trade_summary <- ag_trade %>% 
  group_by(Country, Element, Unit, Year) %>% 
  summarise(sum_value = sum(Value, na.rm=TRUE)) %>% 
  mutate(Element_group = ifelse(Element %in% quantity_vals, 'By Quantity', 'By Value'))

ag_trade_summary$Element <- factor(ag_trade_summary$Element, levels = 
                                           c('Export Quantity', 'Import Quantity',
                                             'Export Value', 'Import Value'))


```

```{r fig.height=6.2}
fig_3a <- ag_trade_summary %>% 
  filter(Element %in% quantity_vals) %>% 
  ggplot(.) + 
  geom_line(aes(x=Year,
                y=sum_value / 1000000,
                group=Element,
                color=Element), size=1.25) + 
  ggtitle('Uganda agricultural imports and exports by quantity and volume') + 
  labs(x='',
       y='Million tonnes\n') + 
  facet_wrap(~Element_group) + 
  scale_x_continuous(breaks = c(1960, 1970, 1980, 1990, 2000, 2010)) + 
  pardee.theme + 
  theme(axis.text.y = element_text(size=16))

ggsave(".//graphics/fig_3a.pdf", plot = fig_3a, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

```{r fig.height=6.2}
fig_3b <- ag_trade_summary %>% 
  filter(Element %in% value_vals) %>% 
  ggplot(aes(x=Year,
                y=sum_value / 1000)) + 
  geom_line(aes(group=Element,
                color=Element), size=1.25) + 
  labs(x='',
       y='Million USD\n',
       caption='Source: FAOSTAT') + 
  facet_wrap(~Element_group) + 
  scale_x_continuous(breaks = c(1960, 1970, 1980, 1990, 2000, 2010)) + 
  scale_y_continuous(labels = comma) + 
  pardee.theme + 
  theme(axis.text.y = element_text(size=12.5))

ggsave(".//graphics/fig_3b.pdf", plot = fig_3b, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

```{r write_fig_3, include=TRUE}
write.csv(ag_trade_summary, './/output_tables/ug_fig_3.csv')
```

```{r ag_trade_detail, include=TRUE}
ag_trade_detail <- readxl::read_xlsx('.//data/ag_trade.xlsx') %>% 
  mutate(Element_group = ifelse(Element %in% quantity_vals, 'By Quantity', 'By Value'),
         IsCoffee = ifelse(str_detect(Item, 'Coffee'), 'Coffee', 'Other')) %>% 
  group_by(Country, Domain, Element, Element_group, Item, IsCoffee, Unit, Year) %>% 
  summarise(sum_value = sum(Value, na.rm=TRUE))
```

```{r fig.height=6.2}
ag_trade_detail %>% 
  filter(Country == 'Uganda',
         Element == 'Export Quantity',
         IsCoffee == 'Coffee') %>% 
  group_by(Country, Year, Element, Item) %>% 
  summarise(sum_value = sum(sum_value, na.rm=TRUE)) %>% 
  ggplot(.) + 
  geom_area(aes(x=Year,
                y=sum_value,
                group=Item,
                fill=Item)) + 
  pardee.theme
```


```{r}
ag_trade_detail %>% 
  ungroup() %>% 
  filter(Element == 'Export Value') %>% 
  group_by(Country, Year, Element, Item) %>% 
  summarise(sum_value = sum(sum_value, na.rm=TRUE)) %>% 
  filter(Year >= 2010) %>% 
  arrange(desc(sum_value))
```

```{r}
ug_trade <- ag_trade_detail %>% 
  ungroup() %>% 
  filter(Country == 'Uganda',
         Element == 'Export Value') %>% 
  spread(Element, sum_value) %>% 
  select(Country, Year, Item, Unit, `Export Value`)

ug_trade_2 <- ag_trade_detail %>% 
  ungroup() %>% 
  filter(Country == 'Uganda',
         Element == 'Export Quantity') %>% 
  spread(Element, sum_value) %>% 
  select(Country, Year, Item, Unit, `Export Quantity`)

ug_trade_3 <- ag_trade_detail %>% 
  ungroup() %>% 
  filter(Country == 'Uganda',
         Element == 'Import Value') %>% 
  spread(Element, sum_value) %>% 
  select(Country, Year, Item, Unit, `Import Value`)

ug_trade_4 <- ag_trade_detail %>% 
  ungroup() %>% 
  filter(Country == 'Uganda',
         Element == 'Import Quantity') %>% 
  spread(Element, sum_value) %>% 
  select(Country, Year, Item, Unit, `Import Quantity`)

#boss join use purrr::reduce()
ug_trade_tables <- list(ug_trade, ug_trade_2, ug_trade_3, ug_trade_4)
ug_trade_master <- reduce(ug_trade_tables, left_join, by=c("Country", "Year", "Item"))

ug_trade_master <- ug_trade_master %>% 
  select(1:3, 5, 7, 9, 11) %>% 
  mutate(net_imports_quantity = `Import Quantity` - `Export Quantity`, 
         price_per_volume = ifelse(`Export Quantity` == 0, 0, `Export Value` / `Export Quantity`))
```

```{r fig.height=6.2}
ug_trade_master %>% 
  filter(Year == 2013) %>% 
  ggplot(.) + 
  geom_point(aes(x=`Export Value`,
                 y=net_imports_quantity,
                 size=price_per_volume)) + 
  ggrepel::geom_text_repel(data=ug_trade_master %>% filter(Year == 2013, abs(net_imports_quantity) >= 150000),
                            aes(x=`Export Value`,
                                y=net_imports_quantity,
                                label=Item), family="Gill Sans MT", size=4.5, force = 24) + 
  #geom_abline(slope = 1, intercept = 0, linetype='dashed') + 
  pardee.theme
```



```{r fig.height=6.2}
fig_4 <- ag_import_dep %>% 
  filter(year >=2018) %>% 
  ggplot(.) + 
  geom_point(aes(x=year,
                 y=val / 100,
                 color=scenario, shape=scenario), size=2.5) + 
  geom_line(aes(x=year,
                y=val / 100,
                color=scenario, group=scenario), size=1.25) + 
  ggtitle('Crop import dependence in Uganda by scenario') + 
  labs(x='Year',
       y='Net imports as percent of demand\n',
       caption='Source: IFs v7.33') + 
  pardee.theme + 
  scale_y_continuous(labels = percent) + 
  scale_color_manual(values = viz_cols) + 
  theme(axis.text.y = element_text(size=16))

ggsave(".//graphics/fig_4.pdf", plot = fig_4, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

```{r write_fig_4, include=TRUE}
write.csv(ag_import_dep, './/output_tables/ug_fig_4.csv')
```

```{r fig.height=6.2}
fig_5 <- maln_pop %>% 
  filter(year >=2018) %>% 
  ggplot(.) + 
  geom_point(aes(x=year,
                 y=val,
                 color=scenario, shape=scenario), size=2.5) + 
  geom_line(aes(x=year,
                y=val,
                color=scenario, group=scenario), size=1.25) + 
  ggtitle('Total malnourished people in Uganda by scenario') + 
  labs(x='Year',
       y='Million people',
       caption='Source: IFs v7.33') + 
  pardee.theme + 
  scale_color_manual(values = viz_cols) + 
  theme(axis.text.y = element_text(size=16))

ggsave(".//graphics/fig_5.pdf", plot = fig_5, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

```{r write_fig_5, include=TRUE}
write.csv(maln_pop, './/output_tables/ug_fig_5.csv')
```

```{r fig.height=6.2}
fig_6 <- yyl %>% 
  filter(year >=2018) %>% 
  ggplot(.) + 
  geom_point(aes(x=year,
                 y=val,
                 color=scenario, shape=scenario), size=2.5) + 
  geom_line(aes(x=year,
                y=val,
                color=scenario, group=scenario), size=1.25) + 
  ggtitle('Total years of life lost to communicable diseases in Uganda by scenario') + 
  labs(x='Year',
       y='Million years',
       caption='Source: IFs v7.33') + 
  pardee.theme + 
  scale_color_manual(values = viz_cols) + 
  theme(axis.text.y = element_text(size=16))

ggsave(".//graphics/fig_6.pdf", plot = fig_6, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

```{r write_fig_6, include=TRUE}
write.csv(yyl, './/output_tables/ug_fig_6.csv')
```

```{r fig.height=6.2}
cum_gdp %>% 
  filter(scenario != 'Current Path') %>% 
  ggplot(.) + 
  geom_point(aes(x=year,
                 y=val,
                 color=scenario, shape=scenario), size=2.5) + 
  geom_line(aes(x=year,
                y=val,
                color=scenario, group=scenario), size=1.25) + 
  ggtitle('Cumulative increase in GDP by scenario (relative to the Current Path)',
          subtitle = '3% discount rate applied') + 
  labs(x='Year',
       y='Billion 2011 USD\n',
       caption='Source: IFs v7.33') + 
  pardee.theme + 
  scale_color_manual(values = viz_cols) + 
  theme(axis.text.y = element_text(size=16))
```


```{r mapping_theme, include=TRUE, echo=TRUE}
#Creating mapping theme
usaid.theme.map <- theme(
  plot.title = element_text(family="Gill Sans MT", color="black", face="bold", size=20, hjust=.5), 
  plot.subtitle = element_text(family="Gill Sans MT", color="black", size=14, hjust=.5), 
  axis.title = element_blank(),
  axis.text = element_blank(),
  axis.ticks = element_blank(), 
  axis.line = element_blank(),
  plot.caption = element_text(family="Gill Sans MT", color="black", size=12), 
  panel.background =  element_blank(), 
  legend.title = element_text(family="Gill Sans MT", color="black", size=14, hjust=.5),
  legend.text = element_text(family="Gill Sans MT", color="black", size=12, hjust=0))
```

```{r mapping_setup, include=TRUE, echo=TRUE}
#I have not found a mapping package that includes the 116 district alignment. So, we need to read in a 
#shapefile (and its associated files) that has the 116 alignment.

shape <- rgdal::readOGR("data", "Uganda_New")
#fortify will take the shape object, convert it to a ggplot readable format, and set the field that has the geo name
shape.fort <- fortify(shape, region = "DNAME_06")
```

# Literacy parity test 

```{r message=FALSE, warning=FALSE}
literacy <- readr::read_csv('.//data/uganda_admin_profile_3.csv') %>% 
  filter(grepl("^3:", Category)) %>% 
  mutate(Literacy_rate = 100-Percent) %>% 
  select(District, Measure, Literacy_rate) %>% 
  spread(Measure, Literacy_rate) %>% 
  select(1:3, 6) %>% 
  rename('Female' = 'Females aged 18 years and above who are illiterate',
         'Male' = 'Males aged 18 years and above who are illiterate',
         'Total' = 'Persons aged 18 years and above who are illiterate') %>% 
  mutate(Literacy_ratio = Female / Male,
         District = ifelse(District == 'Kampala-Kcca', 'Kampala',
                    ifelse(District == 'Mityana1', 'Mityana', as.character(District))))
```

```{r}
regions <- read.csv('.//data/regions.csv')
```

```{r}
literacy <- literacy %>% 
  left_join(x=.,
            y=regions,
            by='District') %>% 
  mutate(Karamoja = ifelse(Subregion == 'Karamoja', 'Karamoja', 'Rest of Uganda'))
```

```{r fig.height=6.2}
literacy %>% 
  ggplot(.) + 
  geom_point(aes(x=Male / 100,
                 y=Female / 100,
                 color=Karamoja), size=2) + 
  geom_abline(slope = 1, intercept = 0, linetype='dashed') + 
  scale_x_continuous(labels = percent,
                     limits = c(0, 1)) + 
  scale_y_continuous(labels = percent,
                     limits = c(0, 1)) + 
  labs(x='Male literacy rate (%)',
       y='Female literacy rate (%)') + 
  pardee.theme + 
  scale_color_manual(values = c('Karamoja' = 'red', 'Rest of Uganda' = 'black'))
```

```{r literacy_map, include=TRUE}
literacy_map <- literacy %>% 
  full_join(x=., 
            y=shape.fort, 
            by=c('District' = 'id'))
```

```{r literacy_map_2, include=TRUE, echo=TRUE, fig.height=7}
literacy_map %>% 
  ggplot(.) + 
  geom_polygon(mapping=aes(x=long,
                           y=lat,
                           group=group,
                           #fill= is the variable you want to map
                           fill=Literacy_ratio),
            #color= sets the boundary color
            color='black', 
            size=.2) + 
  ggtitle('Ratio of Female to Male literacy rates in Uganda') + 
  coord_map() + 
  labs(fill='') + 
   #    caption='Source: 2014 data from LandScan\n2015 data from the Global Human Settlements Layer') + 
  #this is how you can change the color scale easily
  #scale_fill_manual(na.value = 'gray50') + 
  usaid.theme.map + 
  theme(legend.position = 'top',
        legend.justification = 'center')
```

# Less than 2 meals per day 

```{r}
hunger_district <- readr::read_csv('.//data/uganda_admin_profile_3.csv') %>% 
  filter(grepl("^14:", Category)) %>% 
  select(District, Measure, Percent) %>% 
  spread(Measure, Percent) %>% 
  rename('Subsistence_prev' = 'Households that depend on subsistence farming as a main source of livelihood',
         'LT_2_meals_daily' = 'Households where (members aged 5 years and above) consume less than two meals in a day') %>% 
  select(District, Subsistence_prev, LT_2_meals_daily) %>% 
  mutate(District = ifelse(District == 'Kampala-Kcca', 'Kampala',
                    ifelse(District == 'Mityana1', 'Mityana', as.character(District))))
  
hunger_district <- hunger_district %>% 
  left_join(x=.,
            y=regions,
            by='District') %>% 
  mutate(Karamoja = ifelse(Subregion == 'Karamoja', 'Karamoja', 'Rest of Uganda'),
         Northern = ifelse(Region == 'Northern', 'Northern', 'Rest of Uganda'))
```

```{r fig.height=6.2}
hunger_district %>% 
  ggplot(.) + 
  geom_point(aes(x=Subsistence_prev / 100,
                 y=LT_2_meals_daily / 100,
                 color=Northern), size=2) + 
  scale_x_continuous(labels = percent,
                     limits = c(0, 1)) + 
  scale_y_continuous(labels = percent,
                     limits = c(0, 1)) + 
  labs(x='Prevalence of subsistence farming as main source of livelihood',
       y='Prevalence of households consuming less than 2 meals daily') + 
  pardee.theme + 
  scale_color_manual(values = c('Northern' = 'red', 'Rest of Uganda' = 'black'))
```

```{r fig.height=6.2}
hunger_district %>% 
  filter(!is.na(LT_2_meals_daily)) %>% 
  #arrange(-LT_2_meals_daily) %>% 
  #top_n(20, LT_2_meals_daily) %>% 
  ggplot(.) + 
  geom_col(aes(x=reorder(District, LT_2_meals_daily),
               y=LT_2_meals_daily / 100,
           fill=Northern),
           color='gray40') + 
  scale_y_continuous(labels = percent) + 
  labs(x='District',
       y='Households consuming less than 2 meals per day',
       caption='Source: UBOS Area Specific Profiles') + 
  coord_flip() + 
  scale_fill_manual(values = c('Northern' = 'red', 'Rest of Uganda' = 'darkgray')) + 
  pardee.theme + 
  theme(axis.text.y = element_blank())
```

# Relative agricultural demand increase 

```{r}
reg_countries <- c('Uganda', 'Kenya', 'Rwanda', 'Tanzania')

rel_agdem_inc <- readxl::read_xlsx('.//data/rel_agdem_inc.xlsx', sheet = 'rel_agdem_inc') %>% 
  gather(year, agdem, 2:37) %>% 
  mutate(Uganda = ifelse(Country == 'Uganda', 'Uganda', 'Other SSA Country'),
         year = as.numeric(year)) %>% 
  mutate(Uganda = factor(Uganda, levels=c('Uganda', 'Other SSA Country'))) %>% 
  mutate(cntry_spotlight = ifelse(Country %in% reg_countries, as.character(Country), 'Other SSA Country')) %>% 
  mutate(cntry_spotlight = factor(cntry_spotlight, levels=c('Uganda', 'Kenya', 'Rwanda', 
                                                            'Tanzania', 'Other SSA Country'))) %>% 
  mutate(cntry_size_ggplot = ifelse(cntry_spotlight == 'Other SSA Country', 'Other SSA Country', 'Spotlight'))

write.csv(rel_agdem_inc, './/output_tables/ug_fig_1.csv')
```

```{r fig.height=6.2}
fig_1 <- ggplot() + 
  geom_line(data=rel_agdem_inc,
            aes(x=year,
                y=agdem,
                group=Country, 
                color=cntry_spotlight, 
                size=cntry_spotlight)) + 
  ggtitle('Increase in total food demand, relative to 2015 values',
          subtitle = 'Uganda vs. other Sub-Saharan African countries') + 
  labs(x='Year',
       y='Relative increase in total food demand from 2015\n',
       caption='Source: IFs Current Path v7.33') + 
  scale_color_manual(values = c('Uganda' = 'darkgreen', 'Kenya' = 'darkblue', 
                                'Rwanda' = 'gold4', 'Tanzania' = 'purple', 
                                'Other SSA Country' = 'gray80')) + 
  scale_size_manual(values=c('Uganda' = 2, 'Kenya' = 2, 'Rwanda' = 2,
                             'Tanzania' = 2, 'Other SSA Country' = .6)) + 
  scale_y_continuous(limits = c(1,4)) + 
  #scale_x_continuous(breaks = c('2015', '2020', '2025', '2030', '2035', '2040', '2045', '2050')) + 
  pardee.theme + 
  theme(axis.text.y = element_text(size=18))

ggsave(".//graphics/fig_1.pdf", plot = fig_1, height = 6.2, width = 10, family="Gill Sans MT", device=cairo_pdf)
```

# Tables

```{r}
scenarios_compare <- readxl::read_xlsx('.//data/results_tables.xlsx', sheet = 'Scenarios compare') %>% 
  gather(scenario, value, 3:6) %>% 
  separate(scenario, into=c("scenario", "year"), sep="_") %>% 
  mutate(scenario = ifelse(scenario == '2018', '2018 value', 
                    ifelse(scenario == 'Current Path', 'Current Path 2050',
                    ifelse(scenario == 'Agricultural Production', 'Agricultural Production 2050',
                           'Food Secure 2050'))),
         category = factor(category, levels = c('Availability', 'Access', 'Stability', 'Utilization')))
  
scenarios_compare <- scenarios_compare %>% 
  mutate(scenario = factor(scenario, levels = c('2018 value', 'Current Path 2050', 
                                                'Agricultural Production 2050', 'Food Secure 2050')))

scenarios_compare$scenario <- str_wrap(scenarios_compare$scenario, width = 12) 

scenarios_compare %>% write.csv('.//output_tables/scenario_compare.csv')
```

```{r fig.height=6.2}
scenarios_compare %>% 
  ggplot(.) + 
  geom_tile(aes(x=scenario,
                y=indicator,
                fill=value), color='gray90') + 
  geom_text(aes(x=scenario,
                y=indicator,
                label = round(value, 1)), color="white", size=6) + 
  facet_grid(category~scenario, scales = "free", space = "free",
             labeller = labeller(groupwrap = label_wrap_gen(width = 6, multi_line = TRUE))) + 
  #ggtitle('Share of population with access to infrastructure types') + 
  #coord_flip() + 
  labs(x='',
       y='') + 
  ggtitle('Trends in food security indicators to 2050 by scenario') + 
  pardee.theme + 
  theme(legend.position = 'none',
        axis.text.x = element_blank(),
        axis.ticks.x = element_blank(),
        #strip.text = element_text(margin = 10),
        strip.background = element_blank()) + 
  scale_fill_continuous(low = 'lightgray', high = 'green4') + 
  scale_y_discrete(labels = function(x) str_wrap(x, width = 25)) + 
  scale_x_discrete(labels = function(x) str_wrap(x, width = 15))
```

```{r}
indicators <- readxl::read_xlsx('.//data/results_tables.xlsx', sheet = 'Indicators') %>% 
  gather(country, value, 3:8) %>% 
  separate(country, into=c("Country", "Year"), sep="_")

#indicators %>% write.csv('.//output_tables/indicators.csv')
```

```{r FAO, include=TRUE}
FAO <- readxl::read_xlsx('.//data/Copy of FAOSTAT_R&UProduction.xlsx', sheet = 'FAOSTAT_R&UProduction') %>% 
  select(-c(`Area Code`, `Element Code`, `Item Code`, `Flag Description`, helper, Flag, Unit)) %>% 
  #unite("Item_Element", c(Item, Element), sep = "_")
  spread(Element, Value) %>% 
  mutate(yield_calc = round(Production / `Area harvested` * 10000, 0),
         yield_test = Yield - yield_calc)
```

```{r top_6_produced_crops, include=TRUE}
top_6_produced_crops <- FAO %>% 
  filter(Year == 2016,
         Area == 'Uganda') %>% 
  #select(Item, Production) %>% 
  top_n(6, Production) %>% 
  select(Item)
  
top_6_produced_crops <- unlist(top_6_produced_crops[,1])
```

```{r crop_cats, include=TRUE}
cash <- c('Coffee, green', 'Cottonseed', 'Tea', 'Tobacco, unmanufactured')

staple <- c('Millet', 'Maize', 'Sorghum', 'Beans, dry', 'Cassava', 
            'Plantains and others', 'Sweet potatoes')

special_crops <- c('Sugar cane', 'Cassava')
```

```{r fig.height=6.2}
FAO %>% 
  filter(Item %in% staple) %>% 
  filter(Year >= 1990) %>% 
  ggplot(.) + 
  geom_point(data=. %>% filter(Year == 1990),
             aes(x=Yield/100000,
                y=Year), color="lightgray", size=4) + 
  geom_line(aes(x=Yield/100000,
                y=Year,  
                color=Item), size=1.2) + 
  geom_point(data=. %>% filter(Year == 2016),
             aes(x=Yield/100000,
                y=Year, color=Item), size=4) + 
  ggtitle('Change in yield of 5 major crops in Uganda and Rwanda',
          subtitle = '1990 - 2016') + 
  labs(caption='Source: FAOSTAT') + 
  facet_wrap(~Area, ncol=1) + 
  #coord_flip() + 
  pardee.theme

```

```{r fig.height=6.2}
FAO <- FAO %>% 
  mutate(Decade = ifelse(Year <1970, '1960s',
                  ifelse(Year >=1970 & Year < 1980, '1970s', 
                  ifelse(Year >=1980 & Year < 1990, '1980s', 
                  ifelse(Year >=1990 & Year < 2000, '1990s', 
                  ifelse(Year >=2000 & Year < 2010, '2000s', '2010s'))))),
         Period = ifelse(Year >=1985, '1960 - 1985',
                  ifelse(Year >1985 & Year <=2010, '1986 - 2010', '2011 - 2016')))


decade_cols <- c('1960s' = "#f7f7f7",
                 '1970s' = "#d9d9d9",
                 '1980s' = "#bdbdbd",
                 '1990s' = "#969696",
                 '2000s' = "#636363",
                 '2010s' = "#252525",
                 '1960 - 1985' = "#f0f0f0",
                 '1986 - 2010' = "#bdbdbd",
                 '2011 - 2016' = "#636363")

FAO %>% 
  filter(Item %in% staple) %>% 
  ggplot(.) + 
  geom_point(aes(x=Production, 
                 y=`Area harvested`, 
                 color=Decade)) + 
  facet_wrap(~Area) + 
  pardee.theme + 
  scale_color_manual(values = decade_cols)
```

# Bilateral trade viz experimentation

```{r}
library(migest)
library(circlize)
```

```{r bilateral, include=TRUE}
#Data is missing.......
bilateral <- read_csv('.//data/FAOSTAT_UganaBilatTrade.csv')
```

```{r}
bilateral %>% 
  distinct(`Reporter Countries`)
```

```{r}
region_groups_fao <- read_csv('.//data/region_groups_fao.csv')
```

```{r}
ug_wheat_trade <- bilateral %>% 
  filter(`Reporter Countries` == 'Uganda' | `Partner Countries` == 'Uganda',
         Item == 'Wheat',
         Element %in% quantity_vals,
         Year == 2013) %>% 
  arrange(desc(Value)) %>% 
  left_join(x=.,
            y=region_groups_fao,
            by=c("Partner Countries" = "FAO_Partner_Countries"))
```

```{r}
ug_wheat_trade %>% 
  arrange()
```

```{r}
library(ggalluvial)
```

```{r fig.height=6.2}
drop_subr <- c('East & Southeast Asia', 'Latin America & the Caribbean', 
               'Middle East & Northern Africa', 'North America', 'Rest of Sub-Saharan Africa')

ug_wheat_trade %>% 
  mutate(new_subgroup = ifelse(sub_group %in% drop_subr, 'Rest of World', as.character(sub_group))) %>% 
  group_by(`Reporter Countries`, `Partner Countries`, new_subgroup) %>% 
  summarise(Value = sum(Value, na.rm = TRUE)) %>% 
  ggplot(aes(axis1 = new_subgroup,
             axis2 = as.factor(`Reporter Countries`),
             y=Value)) + 
  scale_x_discrete(limits = c('Trade Partner', 'Uganda'), expand = c(.1, .05)) + 
  ggtitle('Wheat imports by source in 2013') + 
  geom_alluvium(aes(fill=`Partner Countries`)) + 
  geom_stratum() + 
  geom_text(stat='stratum', label.strata = TRUE, fontface="bold") + 
  pardee.theme + 
  theme(legend.position = 'none')
```


```{r}
bilat_staple <- bilateral %>% 
  filter(`Reporter Countries` == 'Uganda' | `Partner Countries` == 'Uganda',
         Item %in% staple,
         Element %in% quantity_vals,
         Year == 2013) %>% 
  arrange(desc(Value)) %>% 
  left_join(x=.,
            y=region_groups_fao,
            by=c("Partner Countries" = "FAO_Partner_Countries"))
```

```{r}
bilat_cash <- bilateral %>% 
  filter(`Reporter Countries` == 'Uganda' | `Partner Countries` == 'Uganda',
         Item %in% cash,
         Element %in% quantity_vals,
         Year == 2013) %>% 
  arrange(desc(Value)) %>% 
  left_join(x=.,
            y=region_groups_fao,
            by=c("Partner Countries" = "FAO_Partner_Countries"))
```


```{r fig.height=6.2}
item_drop <- c('Millet', 'Sweet potatoes')

bilat_staple <- bilat_staple %>% 
  filter(!Item %in% item_drop,
         `Partner Countries` != 'Unspecified Area') #very little trade in these

drop_subr <- c('East & Southeast Asia', 'Latin America & the Caribbean', 
               'Middle East & Northern Africa', 'North America', 'Europe', 
               'Oceania', 'South Asia')

bilat_staple %>% 
  mutate(new_subgroup = ifelse(sub_group %in% drop_subr, 'Rest of World', as.character(sub_group))) %>% 
  group_by(`Reporter Countries`, `Partner Countries`, Item, new_subgroup) %>% 
  summarise(Value = sum(Value, na.rm = TRUE)) %>% 
  filter(!is.na(new_subgroup)) %>% 
  ggplot(aes(axis1 = new_subgroup,
             axis2 = as.factor(`Reporter Countries`),
             y=Value)) + 
  scale_x_discrete(limits = c('Trade Partner', 'Uganda'), expand = c(.1, .05)) + 
  ggtitle('Trade in staple crops to Uganda in 2013 by source region') + 
  geom_alluvium(aes(fill=`Partner Countries`)) + 
  geom_stratum() + 
  geom_text(stat='stratum', label.strata = TRUE, fontface="bold") + 
  pardee.theme + 
  facet_wrap(~Item, nrow=2) 
  theme(legend.position = 'none')
```

```{r}
library(FAOSTAT)

?getFAO

obj1 <- FAOSTAT::getFAO(name = NULL,
                        domainCode = 'TM', 
                        elementCode = 5312, 
                        itemCode = 56,
                        countrySet = '226',
                        useCHMT = TRUE,
                        yearRange = c(2000, 2016),
                        outputFormat = 'long')


```



```{r}
FAO_bilateral_staple <- read_csv('FAOSTAT_data_7-12-2018.csv') %>% 
  left_join(x=.,
            y=region_groups_fao,
            by=c("Partner Countries" = "FAO_Partner_Countries")) %>% 
  rename('partner_sub_group' = 'sub_group',
         'partner_sub_group_eac' = 'sub_group_eac') %>% 
  select(-c(main_group, Country)) %>% 
  left_join(x=.,
            y=region_groups_fao,
            by=c("Reporter Countries" = "FAO_Partner_Countries")) %>% 
  rename('reporter_sub_group' = 'sub_group',
         'reporter_sub_group_eac' = 'sub_group_eac') %>% 
  select(-c(main_group, Country))
```

```{r fig.height=6.2}
FAO_bilateral_staple %>% 
  filter(partner_sub_group_eac == 'East African Community') %>% 
  #mutate(new_subgroup = ifelse(sub_group %in% drop_subr, 'Rest of World', as.character(sub_group))) %>% 
  #group_by(`Reporter Countries`, `Partner Countries`, Item, new_subgroup) %>% 
  #summarise(Value = sum(Value, na.rm = TRUE)) %>% 
  #filter(!is.na(new_subgroup)) %>% 
  ggplot(aes(axis1 = partner_sub_group_eac,
             axis2 = reporter_sub_group_eac,
             y=Value)) + 
  scale_x_discrete(limits = c('Trade Partner', 'Reporter Country'), expand = c(.1, .05)) + 
  ggtitle('Regional flow in staple crops by world region in 2016') + 
  geom_alluvium(aes(fill=Value)) + 
  geom_stratum() + 
  geom_text(stat='stratum', label.strata = TRUE, fontface="bold") + 
  pardee.theme + 
  scale_y_continuous(labels = comma) + 
  #facet_wrap(~Item, nrow=2) 
  theme(legend.position = 'none')
```