Unfortunately your stats all look OK (meaning that the problem is not obvious and may be harder to fix). The problem you describe sounds like over-subscription, which is where VM have sold more connections than they have the local capacity to serve at peak times, but I don't recognise the characteristic peak time rise in latency in your ping plot. Doesn't mean it isn't there, but you might want to set up a Thinkbroadband BQM and see what that shows.
Fingers crossed that this is an area problem and there's already a resolution date set, because if not (and it doesn't mysteriously "fix itself) then it could be a pain to get identified and solved.
Do an Ookla speed test on a device connected by an ethernet cable, tell us what that shows from a range of tests? If any tests are below the minimum guaranteed speed that'll give VM the necessary shove to act, whereas they make no guarantees regarding latency.
Unfortunately it's still really bad. Both yesterday (Monday) and now and 8:30 Tuesday evening I've been getting speeds of around 8Mbps whilst I pay for 100!
This seems to now be happening every night from around 6:30 until at least 10:30 when I went to bed.
My friend who lives around the corner from me is seeing the exact same issues at the exact same times every night. I presume the area has been massively oversold and there isn't enough capacity to meet demand? Is there anything that can be done about this?
Here's a screenshot of tonight. It's not even the speed that's annoying, it's the constant spikes up to 400 ping every two or three seconds.