@hobbs has the correct answer - look at the VSWR only, and see where the antenna is working.
The antenna analyser, and any network analyser, measures the magnitude and phase of the reflection coefficient presented at its terminals. It does this at many frequencies. From this it can display a graph of the VSWR, return loss etc.
It can convert to other useful numbers, by calculating the values of components in a made-up equivalent circuit. The simplest circuit is a series R and X, and we often talk about an impedance of 45 + j10 ohms. It's also possible to solve for the values of a parallel R and X, which isn't so intuitive but can be useful for matching purposes. This is what you're displaying. It could be expressed as a series R and L or R and C or parallel RL or RC, the only difference being that it's converted from X to L with the usual formula.
These equivalent circuit transforms are useful when measuring the values of unknown components - I use my NanoVNA all the time to measure inductors and capacitors. You can also use the values to guide you in matching a non-resonant antenna. One big caveat is that the analyser tells you the values right at the analyser port, unless you calibrate "through the coax" or dial in a port extension. The VSWR doesn't change (much) but you know that L turns to C, low impedance turns to high, every quarter wave along the coax, so the figures measured at the analyser have little relation to what's happening at the antenna. For actual antenna tuning, you must use as short a cable as possible, and put your open/short/load standards at the far end of the cable.
To your radio problem - if the VSWR looks about right for the antenna (slightly better because of the long cable) then the connection is probably ok. You can always check the cable by measuring its reflection coefficient with the other end disconnected.
Have you checked that the tone squelch is turned off on your radio?
Have you looked around the band for strong signals nearby?
Very strong external signals may show up on the rigexpert as a "noisy" data point, look for those.