Data: input.csv holds 15 merged spectra (energy 1.196-1.606 eV, 1024 points) for powers 5-100 uW.
Method: each spectrum was fitted with a sum of 3 Lorentzians (peaks near 1.33, 1.54 and an edge peak above 1.6 eV) plus a linear background, with scipy curve_fit. The dominant peak is at ~1.33 eV.
Lorentzian results for the dominant peak (power uW: intensity, xc in eV):
5: 121.7, 1.31789 | 10: 225.3, 1.32299 | 15: 316.0, 1.32491 | 20: 406.0, 1.32656 | 25: 497.3, 1.32793 | 30: 591.3, 1.32896 | 35: 667.0, 1.32946 | 40: 793.7, 1.33074 | 45: 860.0, 1.33112 | 50: 954.0, 1.33210 | 60: 1137.9, 1.33255 | 70: 1337.2, 1.33379 | 80: 1522.0, 1.33400 | 90: 1742.4, 1.33465 | 100: 2016.6, 1.33529
Plots (data saved in plotdata.csv): intensity vs power and xc vs power, each fitted with the allometric law y = a * P^b.
Allometric fit, intensity: a = 19.1 +- 1.5, b = 1.004 +- 0.019 (unweighted fit). The fit weighted by Lorentzian errors gives b = 0.965 +- 0.002, but that error is too small because it ignores model uncertainty.
Allometric fit, peak position: a = 1.3096 +- 0.0004 eV, b = 0.0043 +- 0.0001. The peak position blue-shifts slightly with power and is almost flat on a log scale.
The secondary peak near 1.54 eV gives intensity b = 0.91 +- 0.03.
Note: no image plots were produced because matplotlib is not installed.
The RESULT line below is the intensity exponent of the dominant peak.

RESULT: b = 1.004 ± 0.019
