In joebrew/maragra: An analysis of malaria at the Maragra sugar company, Mozambique

# csl: journal-of-health-economics.csl

# Packages
library(tidyverse)
library(knitr)
library(Hmisc)
library(brew)
library(maragra)
library(RColorBrewer)
library(extrafont)
library(kableExtra)
library(raster)
loadfonts()
## Global options
options(max.print="75")
opts_chunk$set(echo=FALSE,
                 cache=FALSE,
               prompt=FALSE,
               fig.height = 4,
               fig.width = 5,
               tidy=TRUE,
               comment=NA,
               message=FALSE,
               warning=FALSE,
               dev = "cairo_pdf",
               fig.pos="!h",
               fig.align = 'center')
opts_knit$set(width=75)
options(xtable.comment = FALSE)

source('prepare_data.R')

\begin{center} \begin{large}

Charts

\end{large} \end{center}

\vspace{5mm}

\begin{center} \textbf{Overview}
\end{center}

\vspace{5mm} \begin{center} \begin{changemargin}{3cm}{3cm}

This document contains charts from Maragra data, meant for comparison with Xinavane. It was created on February 5, 2019 by Joe Brew for Laia Cirera and Elisa Sicuri.

\end{changemargin} \end{center}

\vspace{20mm}

\noindent\fbox{% \parbox{\textwidth}{% \subsection*{Action} \begin{itemize} \item Review charts herein \item Request clarification or ask questions if applicable \end{itemize} \vspace{2mm} }% }

\vfill \null

\subsection*{Desinataires} \textbf{Laia Cirera; Elisa Sicuri}

\vspace{3mm}

\newpage

Malaria prevalence in the region, (2013-2017), Manhiça (control)

load('xin_bes.RData')

x <- xin_bes %>%
  filter(district %in% c('MAGUDE', 'MANHICA')) %>%
  filter(date >= '2013-01-01',
         date <= '2017-12-31',
         disease == 'MALARIA') %>%
  group_by(month = as.Date(cut(date,'month')),
           district) %>%
  summarise(cases = sum(cases)) %>%
  mutate(year = as.numeric(format(month, '%Y'))) %>%
  left_join(pop %>%
              group_by(year, district) %>%
              summarise(population = sum(population))) %>%
  mutate(p = cases / population * 1000)

ggplot(data = x %>% filter(district == 'MANHICA'),
       aes(x = month,
           y = p)) +
  geom_line() +
  geom_area(fill = 'darkblue', alpha = 0.2) +
  facet_wrap(~district) +
  labs(x = 'Month',
       y = 'Incidence per 1,000',
       title = 'Malaria incidence') +
  databrew::theme_databrew()

Malaria prevalence in the region, (2013-2017)

Magude (intervention)

ggplot(data = x %>% filter(district == 'MAGUDE'),
       aes(x = month,
           y = p)) +
  geom_line() +
  geom_area(fill = 'darkblue', alpha = 0.2) +
  facet_wrap(~district) +
  labs(x = 'Month',
       y = 'Incidence per 1,000',
       title = 'Malaria incidence') +
  databrew::theme_databrew()

Both locations

ggplot(data = x,
       aes(x = month,
           y = p)) +
  geom_line() +
  geom_area(fill = 'darkblue', alpha = 0.2) +
  facet_wrap(~district) +
  labs(x = 'Month',
       y = 'Incidence per 1,000',
       title = 'Malaria incidence') +
  databrew::theme_databrew()

Number of workers by contract type

library(databrew)
plot_data <- ab_panel %>%
  filter(date <= '2016-02-01') %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  mutate(group = department) %>%
  # mutate(field = ifelse(department == 'Field', 
  #                            'Field worker',
  #                            'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(group = permanent_or_temporary, year_month) %>%
  summarise(n = length(unique(oracle_number)))

ggplot(data = plot_data,
       aes(x = year_month,
           y = n,
           color = group)) +
  geom_line() +
  scale_color_manual(name = '',
                     values = databrew::make_colors(n = 3)) +
  theme_databrew() +
  labs(x = 'Month',
       y = 'Workers',
       title = 'Number of workers by month')

Absenteeism rate by contract type

library(databrew)
plot_data <- ab_panel %>%
  filter(date <= '2016-02-01') %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  mutate(group = department) %>%
  # mutate(field = ifelse(department == 'Field', 
  #                            'Field worker',
  #                            'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(group = permanent_or_temporary, year_month) %>%
  summarise(abs = length(which(absent)),
            denom = n()) %>%
  ungroup %>%
  mutate(p = abs / denom * 100)

ggplot(data = plot_data,
       aes(x = year_month,
           y = p,
           color = group)) +
  geom_line() +
  scale_color_manual(name = '',
                     values = databrew::make_colors(n = 3)) +
  theme_databrew() +
  labs(x = 'Month',
       y = '%',
       title = 'Absenteeism rate by month')

Permanent workers absenteeism rate and malaria incidence

plot_data <- ab_panel %>%
  filter(date <= '2016-02-01') %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
    filter(permanent_or_temporary == 'Permanent') %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  mutate(group = department) %>%
  # mutate(field = ifelse(department == 'Field', 
  #                            'Field worker',
  #                            'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(year_month) %>%
  summarise(abs = length(which(absent)),
            denom = n()) %>%
  ungroup %>%
  mutate(p = abs / denom * 100)


x <- xin_bes %>%
  filter(district %in% c('MAGUDE', 'MANHICA')) %>%
  filter(date >= '2013-01-01',
         date <= '2017-12-31',
         disease == 'MALARIA') %>%
  group_by(month = as.Date(cut(date,'month')),
           district) %>%
  summarise(cases = sum(cases)) %>%
  mutate(year = as.numeric(format(month, '%Y'))) %>%
  left_join(pop %>%
              group_by(year, district) %>%
              summarise(population = sum(population))) %>%
  mutate(p = cases / population * 1000)

ggplot(data = plot_data,
       aes(x = year_month,
           y = p * 3)) +
  geom_line(color = 'darkred') +
  theme_databrew() +
  labs(x = 'Month',
       y = '%',
       title = 'Absenteeism rate by month') +
  geom_line(data = x %>% filter(district == 'MANHICA'),
       aes(x = month,
           y = p),
       color = 'blue',
       lty = 2) +
  labs(x = 'Month',
       y = 'Incidence per 1,000',
       title = 'Malaria incidence and permanent worker absenteeism',
       subtitle = 'Red = absenteeism; blue = malaria') +
  databrew::theme_databrew() +
  xlim(as.Date('2013-01-01'),
       as.Date('2016-02-01')) +
  scale_y_continuous(sec.axis = sec_axis(~./3, name = 'Absenteeism'))

Number of temporary workers by job type

(Skipped due to category incompatibility)

Temporary workers absenteeism rate and malaria incidence

(Modified due to category incompatability)

plot_data <- ab_panel %>%
  filter(date <= '2016-02-01') %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
    filter(permanent_or_temporary == 'Temporary') %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  mutate(group = department) %>%
  # mutate(field = ifelse(department == 'Field', 
  #                            'Field worker',
  #                            'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(year_month, group) %>%
  summarise(abs = length(which(absent)),
            denom = n()) %>%
  ungroup %>%
  mutate(p = abs / denom * 100)


x <- xin_bes %>%
  filter(district %in% c('MAGUDE', 'MANHICA')) %>%
  filter(date >= '2013-01-01',
         date <= '2017-12-31',
         disease == 'MALARIA') %>%
  group_by(month = as.Date(cut(date,'month')),
           district) %>%
  summarise(cases = sum(cases)) %>%
  mutate(year = as.numeric(format(month, '%Y'))) %>%
  left_join(pop %>%
              group_by(year, district) %>%
              summarise(population = sum(population))) %>%
  mutate(p = cases / population * 1000)

ggplot(data = plot_data,
       aes(x = year_month,
           y = p * 3)) +
  geom_line(aes(color = group)) +
  theme_databrew() +
  labs(x = 'Month',
       y = '%',
       title = 'Absenteeism rate by month') +
  geom_line(data = x %>% filter(district == 'MANHICA'),
       aes(x = month,
           y = p),
       color = 'blue',
       lty = 2) +
  labs(x = 'Month',
       y = 'Incidence per 1,000',
       title = 'Malaria incidence and permanent worker absenteeism',
       subtitle = 'Blue dotted line= malaria') +
  databrew::theme_databrew() +
  xlim(as.Date('2014-01-01'),
       as.Date('2016-02-01')) +
  scale_y_continuous(sec.axis = sec_axis(~./3, name = 'Absenteeism')) +
  scale_color_manual(name = 'Worker type',
                     values = databrew::make_colors(n = 3))

\newpage

Other miscellaneous charts

Number of workers by department

library(databrew)
plot_data <- ab_panel %>%
  filter(date <= '2016-02-01') %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  mutate(group = department) %>%
  # mutate(field = ifelse(department == 'Field', 
  #                            'Field worker',
  #                            'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(group, year_month) %>%
  summarise(n = length(unique(oracle_number)))

ggplot(data = plot_data,
       aes(x = year_month,
           y = n,
           color = group)) +
  geom_line() +
  scale_color_manual(name = '',
                     values = databrew::make_colors(n = 3)) +
  theme_databrew() +
  labs(x = 'Month',
       y = 'Workers',
       title = 'Number of workers by month')

Number of temporary workers over time

plot_data <- ab_panel %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  # mutate(group = department) %>%
  mutate(group = ifelse(department == 'Field',
                             'Field worker',
                             'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(group, year_month) %>%
  summarise(n = length(unique(oracle_number)))

ggplot(data = plot_data,
       aes(x = year_month,
           y = n,
           color = group)) +
  geom_line()

Absenteeism rate over time

plot_data <- ab_panel %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  mutate(group = department) %>%
  # mutate(group = ifelse(department == 'Field',
  #                            'Field worker',
  #                            'Not field worker')) %>%
  # mutate(group = paste0(permanent_or_temporary, ' ',
  #                         tolower(field))) %>%
  group_by(group, year_month) %>%
  summarise(abs = length(which(absent)),
            eligibles = n()) %>%
  mutate(ab_rate = abs / eligibles * 100)

ggplot(data = plot_data,
       aes(x = year_month,
           y = ab_rate,
           color = group)) +
  geom_line()

Absenteeism rate over time by worker type

plot_data <- ab_panel %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  # mutate(group = department) %>%
  mutate(field = ifelse(department == 'Field',
                             'Field worker',
                             'Not field worker')) %>%
  mutate(group = paste0(permanent_or_temporary, ' ',
                          tolower(field))) %>%
  group_by(group, year_month) %>%
  summarise(abs = length(which(absent)),
            eligibles = n()) %>%
  mutate(ab_rate = abs / eligibles * 100)

ggplot(data = plot_data,
       aes(x = year_month,
           y = ab_rate,
           color = group)) +
  geom_line()

Number of workers over time by worker type

plot_data <- ab_panel %>%
  left_join(workers %>% dplyr::select(oracle_number,
                                      department,
                                      permanent_or_temporary)) %>%
  mutate(year_month = as.Date(paste0(format(date, '%Y-%m'), '-01'))) %>%
  # mutate(group = department) %>%
  mutate(field = ifelse(department == 'Field',
                             'Field worker',
                             'Not field worker')) %>%
  mutate(group = paste0(permanent_or_temporary, ' ',
                          tolower(field))) %>%
  group_by(group, year_month) %>%
  summarise(n = length(unique(oracle_number)))

ggplot(data = plot_data,
       aes(x = year_month,
           y = n,
           color = group)) +
  geom_line()

Bibliography

joebrew/maragra documentation built on Aug. 11, 2020, 8:39 p.m.