In a clever use of copper-clad board, he was able to make a reasonable reflector by soldering together three boards and an RF connector. A single wire makes the “driven element,” and by bending it to just the right position, you can change the characteristic impedance for matching.
The antenna, in this case, is essentially a quarter-wave antenna with a ground plane and reflector arrangement. After the obligatory chalk talk, he breaks out the vector network analyzer and shows how well it matches. He didn’t, however, measure the gain or directional selectivity due to the reflector.