I use a Blue Yeti, and when we record groups in-person that's what we use since it can be easily switched into omnidirectional to catch sound all around it. For most sessions, I'm solo in the room and put it in cardioid to reduce background noise. Tony uses a Blue Snowball on his end (we record through Zencastr). So, if you want a USB solution, we both are satisfied with those models.

I have a little portable recorder (a Tascam DR-05) for events, since it can fit in my pocket and has some decent-sized mics on it for a portable unit, but I've not used any content from it yet for the podcast (I keep forgetting to bring it!).

In terms of background noise, reducing sensitivity on the mic to try and only capture sound near it is the best thing (on the Yeti this is done via a gain control knob; at home I keep it all the way down, but if people are around I have to turn it up the further from the mic they sit).

Beyond that, you can try to clean it up in post, but that works better for pulling out consistent sounds (like a fan humming in the background) versus other people talking.

We used the Yeti to record Keri Wing (first female winner of a PAPA circuit event) for our 10th episode outside in a public area. It serves as a good example of how hard it is to deal with background noise. While the conversation itself is clear, the background noise is also always there (I got a lot of it to the point where you couldn't tell what everyone in the background was saying all the time, but I never was happy with my cleaning effort on it and wish I'd done a better job in both pre- and post-production). On retrospect I should have pulled the gain down even more, but live and learn. Honestly, my recommendation is don't record in loud public places, it is just way too much work to deal with cleaning it up (or accept it and apologize to listeners before playing the clips!).

Do ya'll use Audacity for editing the podcast?

Yes, for sound quality and consistency, there is a free plugin that works great (I use it on the Eclectic Gamers Podcast). It is called Compress Dynamics 1.2.6 and can be downloaded here: https://theaudacitytopodcast.com/chriss-dynamic-compressor-plugin-for-audacity/

When installed it will show in your Effects menu in Audacity. What I do is import all the tracks I am going to be working with, select them all, and run the tool on them all upfront, before I do any other editing. Compressors boost quiet sections, limiters quiet loud sections. This does both tasks, hence why it is so useful if audio is coming in at different volumes.

Now, when you run it, you'll have a menu of five options, and that can be confusing. Here is a list of settings that I've seen recommended for podcasts:
Compress ratio: 1.000
Compression hardness: 0.250 - 0.750 (0.650 is a good starting point)
Floor: -32.00
Noise gate falloff: 0.00
Maximum amplitude: 1.000

Here are the current settings I use for EGP:
Compress ratio: 1.000
Compression hardness: 0.250
Floor: -20.00
Noise gate falloff: 0.00
Maximum amplitude: 1.000

Having your compress ratio at 1 will make the levels very consistent. The maximum amplitude of 1 makes the audio loud (so you don't have to crank the volume up to hear things). These two settings will make the biggest difference for you.

Compression hardness determines the dynamic range. Too low and you'll lose soft sounds. Too high and breathing itself can become distracting (due to being too loud). Obviously, I favor risking the soft sounds, but your mileage may vary.

The floor is the point where you want the tool to treat sound as unwanted. Too high (closer to 0) and you may lose soft sounds, too low (like going to -48.00) and you may keep leftover noise you don't want.

Noise gate falloff will determine how the compressor will deal with the sounds below the floor. The current settings do nothing, they stay where they are. If you were to change the 0 value to, say, 2, then it would reduce the audio by a factor of 2. So, let's say you had the recommended value of -32 as your floor. And let's say you had a sound (like setting down a beer bottle) that came in softly at -33. That means the noise gate falloff would be applied. At 0 that means nothing, but at 2 it would reduce you -33 sound by a factor of 2, making it -66 (this is all dB). So, it makes it quieter.

I've not done this yet, but I think I will on our next episode (using a 2 at the -20 floor I've chosen). I'm trying to get rid of keyboard/mouse clicks without manually editing them out to make our sound a bit cleaner than is current.

Anyway, Compress Dynamics is awesome and I think you'll find it very helpful for improving sound quality with next to no time investment added to your project (you just click after choosing the settings and wait a couple minutes for the audio to process).