ghe-bomad/ecotrace-pilot-deployment-kiambu

# ecoTrace biogas-sensor pilot: installation and feasibility, Kiambu County, Kenya Figures and figure-generating code for the installation and social-science short communication: a feasibility assessment of low-cost, manufacturer-independent digital MRV for household biogas, from the ~6-month, 20-household Kiambu pilot. The paper argues that low-cost independent monitoring can capture operational status and use rhythms but is not yet a turnkey route to calibration-free absolute quantification. The manuscript prose lives in the authors' private paper repository; only the code and the final figures are here. The survey analysis, the cohort use-rhythm analysis and the feasibility check against literature production envelopes belong to the companion technical paper, *ecoTrace biogas-sensor pilot, Kiambu County, Kenya: sensor, calibration and CFD* — cited here by name only, not by repository path. This repository's own `fig_diurnal` is narrower: one household's normalised, shape-only use rhythm, not the cohort analysis. ## Contents | path | what | |---|---| | `make_figures.py` | self-contained generator for `fig_coverage` and `fig_diurnal` | | `scientific.mplstyle` | figure style (LaTeX serif, Okabe-Ito, 600 dpi) | | `figures/` | the rendered figures, including the manually produced site maps | | `scripts/get_data.sh` | clones and derives the companion dataset | ## The figures - **`fig_coverage`** (Figure 1): daily telemetry-completeness heatmap, one row per device serial, de-identified. Reads raw Parquet from the data repository and re-derives what it needs. - **`fig_diurnal`** (Figure 2): one household's normalised daily use rhythm. Each day is normalised to 100%, so the figure is shape-only and carries no absolute volumes. Reads the data repository's derived CSVs. - **`installation_sites_22.png`**: map of the 22 installation sites in Kiambu County. Produced manually, not by code in this repository: household GPS is not part of the published dataset, so the map c …