If I was Keysight I'd contact Nilu and David and ask them to prove where and how they were able to garner so many votes. If it's legit, then proof is easy. "I posted it on Facebook and all my friends voted would be a demonstrably BS answer".
While a statistical analysis and experience might agree that the indication is very clear, is that a sufficiently adequate cause to act as if it is absolutely
proven?
Considering the fact that vote buying is not exactly new, I would hope there are means to track these down.
If it were possible, I'd like to see the MAC addresses (better than IP addresses) of every voting machine collated from multiple competitions, grouped by category of competition, to see if there is a trend. The only way vote sellers could bypass this is to have frequent MAC address changes.
If that's not such a great idea - then what would be?