Analysis Report
Statistical analysis of moss-based eco-cooling system performance — descriptive stats, t-tests, cooling effects, diurnal patterns & more.
Source: environment_history CSV
🔍 Filters
Narrow the analysis by date range, time-of-day window, and humidity range. Humidity filter applies to all humidity sensors (outdoor, near-moss, near-non-moss).
1 Descriptive Statistics
Summary statistics across all sensor channels (computed from raw per-minute data).
| Sensor | Mean | Std Dev | Min | Max |
|---|
2 Time Series Analysis
Hourly-averaged plots showing diurnal cycling and sensor behaviour over time.
⚠ Data Quality Note: The Moss Wall Temp plot shows repeated extreme negative spikes (down to roughly −50 to −130°C) on at least three occasions — not a real thermal signal. These are likely sensor dropouts or wiring faults.
Hourly Humidity Overview
3 Moss vs Non-Moss Comparison (t-tests)
Paired t-test results confirming statistically significant differences.
| Comparison | Moss Mean | Non-Moss Mean | t-stat | p-value | Significance |
|---|
Wall Temp Box Plot
Surface Temp Box Plot
Humidity Box Plot
4 Cooling Effect
Outdoor temp minus wall temp — positive values mean the wall is cooler than outdoors.
⚠ Artifact Alert: The histogram shows a cluster of moss cooling values around 150–160°C, consistent with −127°C floor readings producing artificial "cooling" values (outdoor ~30°C minus wall ~−127°C ≈ 157°C).
Distribution of Cooling Effect
Hourly Moss Cooling Advantage
5 Diurnal Patterns
Daytime (06:00–18:00) vs night-time (18:00–06:00) temperature differences.
| Period | Moss Wall Temp (°C) | Non-Moss Wall Temp (°C) | Diff (°C) |
|---|
Diurnal Temperature Pattern
Hourly Average Temperatures
6 Humidity Buffering
Standard deviation of humidity readings — lower values indicate more stable microclimate.
| Location | Humidity Std Dev (%) | Interpretation |
|---|
Hourly Near-Wall vs Outdoor Humidity
Humidity Variability
7 Correlation Analysis
Pearson correlation heatmap across sensor channels. Near-zero correlations indicate noise-dominated data.
ℹ Note: "Moss Wall" correlates near zero with everything else (0.03 to 0.09, one value −0.04), while every other temperature pair correlates strongly (0.92–0.98). A sensor measuring the same physical process should track them — near-zero correlation is evidence the series is dominated by noise/dropouts.
1.00 (self)
Strong +ve (≥ 0.7)
Moderate +ve (0.4–0.7)
Weak (−0.4 to 0.4)
Moderate −ve (−0.7 to −0.4)
Strong −ve (≤ −0.7)
Correlation Matrix
8 Evapotranspiration Evidence
Statistical evidence linking humidity buffering to temperature reduction.
Evapotranspiration Correlation
Humidity Buffering Index (BI)
Hourly Tracking
9 Key Findings
Summary of major results from the analysis.