Car park data problems

We've recenty been havng some problems with our feeds of car park usage data. Our feed of data for the City Centre car parks wasn't working between 2018-05-03 and 2018-06-04, and seperately our feed of data for the Park & Ride car parks hasn't been working since 2018-05-21 and has just been fixed today (2018-06-12). Both are now working, but we have lost the historical data that would otherwise have been collected during these periods.

Seperately, we've ovserved problems with the data we are recieving for several car parks. The number of spaces reported as used in the P&R Madingley Road and P&R Milton car parks never decreases during the day, and the number of spaces reported as used in the P&R Newmarket Road Front car park seems to stick at the same value for protracted periods. We've reported these to the relavent council.

The city-centre car park feed failed again between 2018-06-18 20:02:05 and 2018-06-19 12:57:04. However this time we managed to investigate and believe that we've implimented a fix.

If you care, it seems that the web server from which we retrieve this data will, sometimes, respond '404 Not Found' to any request that doesn't include an 'Accept-Encoding' header. You'll apreciate this was tricky to identify because the URL in question worked fine when accessed from a browser...

Hi Jon,

There still seem to be problems with: * Madingley P&R exit sensor * Newmarket Rd P&R front sensors * All data periodically, with random zero or off-the-scale readings and discontinuities (e.g. 31 July)

Are there any cases where the original data is still available somewhere and could be re-imported? Would the be worth doing?

For any data that is irretrievably bad, wouldn't it be better to erase it, so that it doesn't distort any future statistical analysis (or, rather, require multiple analysts to figure out how to filter out the bad data)?

Best wishes,

Edward

we've reported the faults with the 'exit' sensors on those car parks and they will be fixed, although we haven't been given a time estimate. I guess the length of the outage has crept up on us - I'll try and get an estimate of how long it will be before the 'loop' is fixed so we can decide whether to suppress the car park completely in the meantime.

At some point we'll 'clean up' the historical data but even just cars 'entering' is potentially useful.

At least P&R Milton does seem to be working again, having apparently been broken since at least December 2016!

The County Council tell us that:

• The sensor loops for the exit counter at Madingley Road have been damaged and they are arranging for these to be re-cut and connected up.
• That Newmarket Road Front P&R car park uses an old radio based system which has difficulty reaching the servers at County Hall and that they are investigating if this can be fixed.

Current status:

• The Newmarket Road Outer sensor seems to have been working much more reliably since 14 August.
• It looks as though the exit counter at Maddingly Road briefly sprang into life on 17 August, but since then the car park has been registering as empty. This would be consistent with the exit counters having been fixed but the entrance counters breaking at the same time.

I'm continuing to liaise with the council about this.

The Maddingley Road sensons were re-fixed on the evening of Mondy 24 September. I think this means we are currently recieveing correct(-ish) data for all the car parks that we monitor.