Not really my usual turf - I'm more crop circles and MIB than cryptids - but this is actually interesting enough to pull me in.
The railway line theory reminds me of ley lines and the way certain locations just seem to accumulate weirdness regardless of what the phenomenon is. Old infrastructure follows old paths, and old paths sometimes follow something older still. Whether that's geology, magnetism, or something weirder, who knows.
We've got a fair stretch of disused railway cutting near me in Lincolnshire. Never personally clocked anything dogman-shaped, but there have been some genuinely odd crop formations within a mile or two of it over the years. Probably coincidence. Probably.
What I'd want to know before getting too excited about the railway connection:
Are these sightings clustered on active lines or abandoned ones?, What era were the lines built?, Any correlation with specific terrain types along the route - woodland, wetland, etc.?
Because ". Near a railway line". Covers a massive amount of geography. You could link almost anything to railway lines if you're not careful with the data. Classic pattern-matching trap.
That said, I'm not dismissing it either. Anyone actually mapped these properly? Would love to see an overlay rather than just anecdotal clustering. Someone with decent GIS skills could do something useful here.