Faecal sludge characterisation dataset and analysis code, Parakou faecal sludge treatment plant, Benin (n=14)

This deposit contains the complete dataset and analysis code underlying a physico-chemical and geochemical characterisation of faecal sludge (boues de vidange, BV) at the faecal-sludge treatment plant (STEP) of Parakou, Benin. It supports a manuscript currently under review. The study has two phases. Phase 1 (7 May to 12 August 2021) is an operational record of all 29 discharge events received at the site, transcribed from the site dumping sheets. Phase 2 (March 2022) comprises 14 daily composite samples collected at the site inlet and characterised for 22 physico-chemical parameters spanning general physico-chemistry (pH, electrical conductivity, temperature, redox potential, total dissolved solids, salinity, dissolved oxygen and saturation), organic load (suspended solids, COD, BOD5 and the COD/BOD5 ratio), nitrogen and phosphorus speciation (nitrate, ammonium, nitrite, orthophosphate), total nutrients (total nitrogen and total phosphorus), and heavy metals (lead and zinc, determined after acid digestion by flame atomic absorption spectrometry, FAAS, Varian SpectrAA 210, AFNOR FD T 90-112, expressed on a dry-matter basis). Cadmium and mercury were not part of the trace-metal suite for this campaign. The values retained correspond to the primary laboratory bench records. During assembly from two complementary source files collected at the same plant, a documented cross-validation was applied and no observation was excluded: two total-dissolved-solids entries were converted from g/L to mg/L after conductivity cross-checking; two total-nitrogen entries were set to the sum of the measured inorganic fractions as a conservative lower bound where the recorded value was physically inconsistent; and lead and zinc were taken from the validated source after a row-shifted corrupted duplicate had been set aside. Borderline values retained without modification are documented. Missing values are coded as NA (maximum missingness: ammonium, 3 of 14 entries). Analytical quality control was limited to five-level calibration curves and method-blank corrections; no analytical replicates, spike-recovery tests or certified reference materials were used. This version provides the complete set of outputs: the ten figures used in the manuscript at 600 dpi (eight main figures plus two supplementary, namely the exploratory multivariate structure and the reduced five-variable PCA robustness check) and the twenty-one result tables produced by the pipeline. Trace-metal indices are reported as geochemical enrichment relative to a crustal reference rather than as pollution; the geoaccumulation index is aggregated as the mean of per-sample values (Igeo Pb 1.76, Igeo Zn 0.71). Figure titles have been removed so that captions live only in the manuscript. The Phase 1 operator-by-consistency association is computed on the real dumping log with a seeded Monte Carlo test (chi-square = 7.25, df = 3, Monte Carlo p = 0.048, B = 10000, set.seed(2024)), the simulation being required because six of the eight expected counts fall below five. Files included (37): BV_Parakou_n14.csv (Phase 2 dataset, 14 rows by 23 columns, UTF-8); phase1_dumping_log.csv (Phase 1 operational log, 29 real events); BV_Base_FINALE_n14_complete.xlsx (working file with the dataset, the correction log and the methods text); BV_Parakou_Analysis.R (self-contained R pipeline reproducing every figure and result table); ten figures at 600 dpi; twenty-one result tables in CSV; README.txt (full documentation, including the variable dictionary for all 23 columns and a list of reproducibility checkpoints); LICENSE (MIT licence text for the code). Licensing: data and documentation are released under CC BY 4.0; the R analysis code is released under the MIT License.