For the linear numbers you can refer to "Photopic lux" or "Melanopic lux" to get the lux value we measured. In the case of the "lights" section, we have normalized most of them to 100lux, but everything else (including "streetlights") is based on actual irradiance as we measured it. When we measure luminance we show both cd/m2 and lux, since we've scaled luminance using our panel model (but this is mostly used for screens).

Our streetlight measurements are a small sample from lights near us in LA. We measured vertical irradiance (looking up standing directly under a light) to avoid stray light from landscape and commercial lighting (so this is the brightest you'd typically encounter).

RP-8 includes "luminance" measurements as roadway cd/m2 specifications too, but that is often to account for the dimmest spot on the roadway, not the brightest. So here we've mostly measured the brightest spot, and I believe diffused illuminance is the best measure of that.

If you have some data you'd like to donate in the meantime, email me a CSV: justgetflux@gmail.com - for now everything is available under our Creative Commons license. We can handle any resolution.