JackPhilpott/MSNA-N-WEC-2026-monitoring

IMPACT Initiatives: MSNA project in Northern Nigeria. This Repo will mainly contain a monitoring RShiny dashboard and cleaning process # MSNA 2026 — Monitoring (cleaning + dashboard) Daily-updating pipeline: pulls survey submissions from KoBo, runs cleaning and diagnostics against the sampling frame (coverage, duplicates, outliers, progress against target per stratum), and updates a Shiny monitoring dashboard. Cleaning and the dashboard run together on the same daily cadence, but live in separate subfolders so the deployable app stays lean. ## Structure - `cleaning/` — KoBo pull + cleaning/diagnostic scripts. Writes its processed output to `data/`. - `dashboard_app/` — The Shiny app itself (`app.R` or `ui.R`/`server.R`). Reads from `data/`. Keep this folder to only what the app needs to run — deployment tools (shinyapps.io, Posit Connect) bundle whatever's in here. - `data/` — Local exchange between cleaning and the dashboard. Gitignored (reproducible from `cleaning/` + `input_data/`). - `input_data/` — Static copy of the sampling frame (from `../1_sampling/output/`) that cleaning diagnostics run against. Gitignored; copy in the file(s) you need, don't reference `1_sampling/` by path. ## Setup ```r renv::restore() # installs the exact package versions from renv.lock ``` `renv.lock` is committed (already tracked); `renv/library/` is gitignored. Open `monitoring.Rproj` in RStudio and renv activates automatically (`.Rprofile` handles this) — no manual `renv::init()` needed, that's already been done. ## Status (2026-08-02) — v1 draft, built against mock data A cleaning script is being built separately and will land within a day or two. To avoid blocking on that, this v1 was built end-to-end against a **simulated submissions dataset** so the dashboard itself could get to ~95% complete before real data exists. Everything below is real except the submission data. - **`input_data/`** populated: WORKING sampling frame (household- and strata-level, 31,051-interview fielding plan, 176 covered LGAs), coverage summary, and admin0/1/2 boundary shapefiles — all copied from `../1_sampling/output/` and ` …