==========================================================================
HOUSEHOLD FOOD SECURITY & WASH — SUMMARY REPORT
Fictional survey · Kobo-to-Report pipeline demo · ALL DATA FICTITIOUS
==========================================================================

BLUF — Households surveyed in five villages show a moderate food
security situation overall, with a concentrated pocket of poor
consumption in the largest village. Open wells remain the main water
source; treatment practices are uneven. Findings below are illustrative
(fictional data) and demonstrate the full chain: collect -> clean ->
analyse -> deliver.

1. KEY FIGURES
   - Households surveyed: 54 (after duplicate removal and cleaning)
   - Food Consumption Score: 4 Poor (7%) · 21 Borderline (39%) · 29 Acceptable (54%)
   - Main water source: Open Well — 20 households (37%)
   - Mean Coping Strategy Index: 6.4 (range 0–12)

2. FOOD SECURITY BY VILLAGE (top of table)
   Koubri       n=13  Poor=1   Borderline=3   Acceptable=9
   Tanghin      n=12  Poor=0   Borderline=4   Acceptable=8
   Ziniare      n=11  Poor=2   Borderline=6   Acceptable=3
   Saaba        n=10  Poor=1   Borderline=3   Acceptable=6
   Loumbila     n=8   Poor=0   Borderline=5   Acceptable=3
   => Attention: Koubri shows the highest share of Poor+Borderline (31% of its households).

3. WASH SNAPSHOT
   - Open Well    20 households (37%)
   - Borehole     16 households (30%)
   - Piped        10 households (19%)
   - River        8 households (15%)
   - No water treatment reported: 21 households

4. METHOD
   - Source: fictional KoboToolbox export (survey_data.csv, 55 raw rows).
   - Cleaning: label standardisation (village casing, water/treatment
     categories), duplicate removal, 'N/A' -> missing, numeric coercion.
   - FCS computed with WFP standard weights (staple x2, pulse x3,
     meat/dairy x4, veg/fruit x1, oil/sugar x0.5); thresholds:
     Poor <=21, Borderline 21.5-35, Acceptable >35.

5. DATA QUALITY NOTES
   - 1 duplicate submission removed (identical household, distinct index).
   - 3 missing values coerced (FCS staple, water treatment, CSI).
   - Village labels required case/whitespace standardisation.

6. LIMITS
   - FICTITIOUS data: no inference for any real location or population.
   - Convenience sample, not representative; no weighting applied.
   - FCS self-reported; no triangulation with consumption recall.

--------------------------------------------------------------------------
Pipeline: kobo_to_report.py -> pivot CSVs -> this report. 
Author: Ahadi Jean Cyrille Dahani — humanitarian analyst, Ouagadougou.
==========================================================================