First, +1 for linking to the datasheet and the schematic.
It depends how you want to use this (reception off air, or cable connection from a transmitter), what accuracy you're expecting, which package the dev board uses, and what the layout of the dev board is.
You'll notice from the datasheet that the real part of the input impedance at 2.4GHz and up is already lower than 50 ohms. Adding a simple shunt resistor will make things worse. A shunt R here would be useful below 1GHz.
You'll also notice that the resistance changes with power level, so you will not achieve a consistent match as your level varies.
However, the input impedance is not far from 50 ohms, and the trajectory of S11 round the Smith chart is fairly tame, so there's nothing exciting going on that might trip you up at a certain frequency. These development boards are built intending to be used with 50 ohm connectors and equipment. Although it's not perfectly matched, simply using with a 50 ohm source will not give you not much level error, maybe a dB or two, and over your extended range from 900M to 5.8G.
If you want to make your dev board 'better' over a broad band and range of input powers, then adding an input pad is probably the easiest way to go about it, if you can afford the drop in level. Adding (for instance) a 6dB pad will only lose you 6dB of input power, and will theoretically improve your S11 by 12dB. This is known in some circles as 'masking' the input impedance. If the layout permits, you could use the R1 site to place a shunt resistor, and then cut the track with a scalpel immediately before and after it to add two series resistors to make a T pad. You should be able to find T pad values with a simple search. However, getting a pad to work well at 5.8G is no mean feat, layout is everything. While at 900MHz, a pad should make things better, at 5.8GHz, it could easily make it worse in practice.