It is nice that it seems doable for ~$200 USD, but the results aren't super encouraging, anyone know of a setup to get better results in a backyard setting?

> Another shot at curve fitting in Excel brought these results closer to expectations, but they were still somewhat off.

Is the raw data available? It would be nice to see the fits and understand why they changed so much. [My unsupported guess is that you must not fit the bigger bump, but a smaller bump on the side.]

The graph has no error bars for the data of the author. I guess the vertical error bar are big...

> Experiments aiming my horn antenna at an Inmarsat geostationary satellite revealed that the angular resolution of my little radio telescope is about 20 degrees.

... and the horizontal error bars are huge.

> Fortunately, you only have to care about the cloud that’s receding the fastest—the one with the largest redshift, in astronomer-speak.

Is it possible to point the device to the other side and analyze the ones with the largest blueshift? This doubles the data points without additional hardware.

The last few are suspiciously accurate. I guess the author is adding the velocity of the Sun/Earth to the data. It would be nice to have that horizontal line in the graph too.