Brilliant! That's exactly what I was unsuccessfully looking for before settling on the scraper. I'd also tried to help with the load problem by restricting to 30 minute increments.

I uploaded v1.0.1 into the same place, now using the XML feed instead.