AnnikaYounge/liberia-processing

# Liberia – code for data processing & analysis The repository contains a reproducible R pipeline for cleaning, matching, and validating pediatric hospital **admission** and **discharge** records in Liberia. The pipeline links admissions to discharges and resolves duplicate cases (readmissions, twins, transfers). The analysis code then provides starting tables and figures about the final dataset. --- ## quick start 1. **install packages** ```r source("R/00_setup.R") ``` 2. **prepare the dataset (cleaning/matching)** ```r source("run_all.R") ``` writes: * `data/processed/final_matched.csv` * `data/processed/unmatched_admissions.csv` * `data/processed/unmatched_discharges.csv` * `data/processed/validation_sample_5pct.csv` 3. **run the analysis (tables + figures)** ```r source("analysis/run_analysis.R") ``` outputs go to `analysis_outputs/`. --- ## input data Place the raw **admission** and **discharge** files in `data/raw/`. * Recommended names: `AdmissionForm.csv`, `DischargeForm.csv`. * From Kobo, use CSV with “_labels” (recommended). XML and excel also works. --- ## repository structure ``` R/ # cleaning + matching 00_setup.R # installs/loads packages 01_data_load.R # read raw data, standardize, basic cleaning 02_matching.R # unique-ID matching 03_matching_nonunique_functions.R # helpers for complex cases 04_matching_nonunique.R # readmissions, twins, transfers; final joins 05_validation.R # builds validation sample analysis/ # analysis (tables + figures) 10_setup.R 20_derive_core.R 30_table1_overview.R 40_time_and_outcomes.R 50_readmissions.R 60_diagnoses_complaints.R 70_malnutrition_labs.R 80_utilization.R 90_geo_and_missing.R run_analysis.R analysis_outputs/ # PNGs + CSVs from the analysis data/ raw/ # put the raw data files here (not versioned) processed/ # final outputs from the pipeline fo …