MAHERYCohortHarmonization

The Madagascar Health and Environmental Research (MAHERY) cohort datasets provide a detailed and interdisciplinary understanding of the relationships between environmental changes, resource use, and human health in Madagascar. These datasets encompass multiple research initiatives, including the Northeast, Antongil, HIARA, and cross-sectional studies, each addressing unique aspects of health and environmental interactions.

• Northeast Dataset: This dataset focuses on communities near Maroantsetra in northeastern Madagascar, where local populations heavily rely on wild foods for nutrition. The study examines the impact of deforestation, biodiversity loss, and hunting on dietary intake, nutritional status, and health outcomes. It includes data on dietary diversity, disease metrics, and environmental resource use[2][3].

• Antongil Dataset: Established in 2015 along Antongil Bay, this dataset evaluates the nutritional value of seafood and the effects of different fisheries governance models—traditional management, co-management, and marine national parks—on fish catch, seafood consumption, and nutritional outcomes. It includes data from 225 households across five communities[4].

• HIARA Dataset: Initiated in 2022 in southwestern Madagascar’s Bay of Ranobe, this study investigates the impacts of artificial coral reef development on fish biomass, fisher livelihoods, nutrition, and mental health. It involves longitudinal monitoring of ecological and social systems across 14 communities[1].

• Cross-Sectional Studies: These studies provide snapshots of dietary patterns, nutritional status, and socio-economic indicators across various regions. They complement the longitudinal datasets by offering broader demographic insights[2][5].

Together, these datasets form a comprehensive resource for understanding how environmental changes influence health outcomes in vulnerable populations.

Sources [1] HIARA study protocol: impacts of artificial coral reef development on … https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2024.1366110/full [2] Cohort Profile: The Madagascar Health and Environmental … https://pmc.ncbi.nlm.nih.gov/articles/PMC5837654/ [3] Seasonal trends of nutrient intake in rainforest communities of north … https://pmc.ncbi.nlm.nih.gov/articles/PMC10260550/ [4] Cohort Description of the Madagascar Health and Environmental … https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2019.00109/full [5] Cohort Profile: The Madagascar Health and Environmental … https://academic.oup.com/ije/article/46/6/1747/3868352 [6] The Madagascar Health and Environmental Research (MAHERY … https://pubmed.ncbi.nlm.nih.gov/29040632/ [7] Integrating approaches to study land use change and hotspots of … https://www.thelancet.com/journals/lanplh/article/PIIS2542-5196(18)30104-9/fulltext [8] Madagascar - MAHAY Study 2016, Endline - Microdata Library https://microdata.worldbank.org/index.php/catalog/study/MDG_2016_MAHAY-EL_v01_M [9] Summary of the PERMANOVA to examine the influence … - Figshare https://figshare.com/articles/dataset/Summary_of_the_PERMANOVA_to_examine_the_influence_of_explanatory_variables_in_the_variation_of_composition_of_coral_assemblages_/21375118

This package establishes a pipeline that harmonizes the MAHERY cohort data across disparate sources and produces a dataset that defines the MAHERY Cohort for future studies.

Approach

We will use the following steps to harmonize the MAHERY cohort data:

1. A comprehensive review of the MAHERY cohort datasets to identify common variables and data structures.
2. Secure identification and procurement of google drive credentials and locations (URLS) of each dataset.
3. Data extraction and cleaning to ensure consistency and compatibility across datasets.
4. Data integration to merge the datasets into a single, harmonized dataset.
5. Data validation to ensure the harmonized dataset is accurate and reliable.
6. Data export to save the harmonized dataset for future studies.

Installation

You can install the development version of MAHERYCohortHarmonization like so:

# FILL THIS IN! HOW CAN PEOPLE INSTALL YOUR DEV PACKAGE?

Pipeline

The pipeline is managed by targets, which is a powerful R package for reproducible data analysis. It allows you to define a series of steps (or targets) that will be executed in order, ensuring that your analysis is reproducible and efficient. The pipeline is deefined in the _targets.R file. The functions we create to build this pipeline are defined in the notebooks in the dev/ directory, and exported to the R directory using fusen.

You can run the pipeline using the following command:

library(targets)

tar_make()

Here is the current state of the pipeline:

library(targets)
library(DiagrammeR)
library(DiagrammeRsvg)
library(rsvg)
#> Linking to librsvg 2.42.7
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
network <- tar_network()
#> -\|/-\|/-\|/-\|/-\|/-\- The project is out-of-sync -- use `renv::status()` for details.
#> |/-\|/-\|/-  checked: 1 | outdated: 0\  checked: 13 | outdated: 0
#> |/ 

# Optional: filter out internal targets like `.Random.seed`
vertices <- network$vertices %>% filter(!grepl("^\\.", name))

# Build node declarations
node_lines <- sprintf('  "%s";', vertices$name)

# Filter edges to match cleaned nodes
edges <- network$edges %>%
  filter(from %in% vertices$name, to %in% vertices$name)

# Build edge connections
edge_lines <- sprintf('  "%s" -> "%s";', edges$from, edges$to)

# Assemble DOT graph string
dot_code <- c("digraph targets_pipeline {", node_lines, edge_lines, "}")
dot_string <- paste(dot_code, collapse = "\n")

# Create grViz graph
graph <- grViz(dot_string)

svg_path <- "man/figures/pipeline_dag.svg"
svg_string <- export_svg(graph)
writeLines(svg_string, svg_path)

Pipeline Graph