Gera Water Service Score (GWSS) — Methodology

Data sources

Formula (step by step)

1. 1

   Download three CSVs from DiscoverWater.co.uk

   Fetch 2025-leakage-per-property-per-day-UK-Current.csv, 2025-Avg-minutes-lost-Current-UK.csv, and 2025-CRI-Current-UK.csv. No API key or account is required — these are plain CSV files served at /csv/<filename>.

2. 2

   Parse and align company data

   Strip "Average" rows. Convert interruption times from HH:MM:SS to decimal minutes (h×60 + m + s/60). Note that some groupings (e.g. "Northumbrian and Essex & Suffolk") are reported as one line — both constituent companies receive the same combined value.

3. 3

   Min-max normalise each metric (inverted: lower=better → higher score)

   For each metric: norm(x) = (max − x) / (max − min) across all 19 companies. This maps the worst performer to 0 and the best to 1. Bounds (2024-25): leakage [64, 169] l/p/d; interruptions [2.0, 141.8] min; CRI [0.00, 10.94].

4. 4

   Compute weighted composite

   GWSS_raw = 0.4 × leakage_norm + 0.3 × interruptions_norm + 0.3 × CRI_norm. This is a number in 0–1.

5. 5

   Scale and round

   GWSS = round(GWSS_raw × 100, 1). Result is in 0–100. There is no second normalisation step — the final 0–100 scale is directly from the weighted 0–1 formula × 100.

Pseudocode

# Input: three DiscoverWater CSVs (no key required)
# Output: GWSS_COMPANIES array

fetch leakage_csv  = "https://discoverwater.co.uk/csv/2025-leakage-per-property-per-day-UK-Current.csv"
fetch interrup_csv = "https://discoverwater.co.uk/csv/2025-Avg-minutes-lost-Current-UK.csv"
fetch cri_csv      = "https://discoverwater.co.uk/csv/2025-CRI-Current-UK.csv"

parse leakage      → { company: litres_per_property_per_day }
parse interruptions → { company: hms_to_minutes(HH:MM:SS) }
parse cri          → { company: cri_float }

# Min-max normalisation (inverted: lower raw → higher norm)
for metric in [leakage, interruptions, cri]:
  mn, mx = min(metric.values()), max(metric.values())
  for each company:
    norm[metric][company] = (mx - raw_value) / (mx - mn)

# Weighted composite
for each company in all_three:
  gwss_raw = 0.4 * norm[leakage][company]
           + 0.3 * norm[interruptions][company]
           + 0.3 * norm[cri][company]
  gwss = round(gwss_raw * 100, 1)   # 0–100

# Grade
if gwss >= 85: grade = "A"
elif gwss >= 70: grade = "B"
elif gwss >= 55: grade = "C"
elif gwss >= 40: grade = "D"
else: grade = "E"

Limitations & caveats

• Company-level, not local: GWSS aggregates each company's full service area. It does not indicate performance at a specific postcode or town. Local performance can vary significantly from the company average.
• CRI reporting groups: Some companies are combined in the DWI CRI data (e.g. Northumbrian & Essex and Suffolk; South Staffs & Cambridge). Constituent companies receive the same combined CRI value.
• Interruptions groupings: Similarly, some company groupings appear in the Ofwat interruptions data. Where companies are combined, both receive the same group value.
• Weighting is Gera's choice: The 40/30/30 split is a judgement call. Users with different priorities (e.g., prioritising water quality over leakage) can reweight the components using the published raw figures.
• Not regulatory guidance: GWSS is a consumer-information index, not a regulatory determination or formal Ofwat assessment. The official performance scorecard is Ofwat's Annual Performance Report, accessible at ofwat.gov.uk.