As a statistician I agree that it is always wise to approach statistics with caution and think about how they have been calculated and whether any bias has been introduced. You are, however, falling into the trap of placing too much emphasis on information that is likely to be biased when you say:
people can now more openly share their experiences, Then there is people who work on services who are openly reporting they have never had to deal with so many crimes/victims/perpetrators (and on limited budgets)
Firstly are the people who work in these services who might be finding that the numbers of crimes/victims/perpetrators is the same or falling as likely to be posting on social media as those who find it's rising and they're struggling? I would expect not.
Secondly, are the numbers of people working in these roles the same, increasing or falling in line with the population? I would suggest that with continuing austerity cuts, it's likely to be falling, which could go some way to explaining why these services are struggling with demand.
I don't mean to suggest that it's not important to listen to the opinions of people who are experiencing issues first hand, that is of course important. What we shouldn't be doing though, is taking these personal experiences and trying to suggest that they indicate a national trend. The statistics from ONS may not be flawless, but the way they are collected will be determined as the best way possible by highly trained professionals and I'd believe those over a general feeling about a situation from reading posts online. (I don't work for the ONS, so have no bias there)