Straybow,
Yep, the UWP lists the best starport there. If you wanted something that was generally useful for others, you could use a traffic-distribution model to programmatically figure out how many starports there are likely to be:
First, figure out the traffic T through the main starport. This would be some percentage of the traffic through the system as a whole -- 50% for example -- or it could be
equal to the official traffic volume through the system. This lets you tune your traffic volume, too.
Next, apply a city-population rule like this:
Starport 1 handles T traffic.
Starport 2 handles T/2 traffic.
Starport 3 handles T/3 traffic.
Starport 4 handles T/4 traffic.
Do that until you reach some threshold traffic number that's too low to bother with, and scatter some number of small starports around to pick up the leftovers.
The article that I'm thinking of which talks about demographics and settlement sizes is here:
"Medieval Demographics Made Easy":
http://www.io.com/~sjohn/demog.htm
For example, a world that has a volume rating of one million passengers per week might actually have starports like this:
Starport 1: 500k passengers/wk
Starport 2: 250k p/wk
Starport 3: 160k p/wk
Starport 4: 125k p/wk
Starport 5: 100k p/wk
Starport 6: 85k p/wk
Starport 7: 70k p/wk
Starport 8: 62k p/wk
Starport 9: 55k p/wk
Starport 10: 50k p/wk
Note that by this point, starports aren't so significant compared to the first few, but still are quite substantial. At 10% we also
might prefer to proceed linearly:
Starport 11: 45k p/wk
Starport 12: 40k p/wk
Starport 13: 35k p/wk
Starport 14: 30k p/wk
Starport 15: 25k p/wk
Starport 16: 20k p/wk
20,000 people per week is about 3000 per day, which is about 120 per hour. That's one 600t liner every 15 minutes. At this point I'm ready to call all this remainder traffic as insignificant, to be subsumed by the ports shown above and smallish ports not worth enumerating.