So here is website is a great resource for actual sales prices: houseprices.io/
You really need a postcode rather than a borough, as the database of addresses doesn't always use the borough name. You also need to interpret the results somewhat - filtering out the "additional price paid transactions" (e.g. lease extensions, or collective enfranchisement of the freehold reversion) from the "standard price paid transactions", and also you need to filter out the sales of leasehold properties where the time interval between sales matches the expected depreciation of a shortened lease term.
But ... what is evident is that there are barely any (and I mean, a handful of instances from amongst the many thousands of transactions) where the nominal price has been lower than the previous one. And in these cases, the reductions are less than 5% which is well within the normal range of variation to be expected from differences in bargaining skill, and certainly not evidence of generalised price falls.
I expect if you looked at those instances in detail, you'd find localised, very specific reasons linked to value destruction that explain those drops, too - cladding issues, remodelling disasters, scenes of violent crime etc. But again, nothing I'd take as symptoms of a generalised price falls.
There's been a fair few cases where the real price (inflation adjusted) has been lower, and these are colour-coded red, but as we've agreed already, this doesn't count as a crash - although any reduction in house prices versus earnings due to wage inflation has the same practical effect for the buyer as a reduction in price - but it won't influence market sentiment and won't cause contagion.
Of course, none of this means that the next set of sales to be uploaded won't be a sea of red with drops in nominal prices too ... but it is completely wrong to claim that the crash has already started when it quite patently hasn't