[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
Maybe best to know what your data-processing criteria are? One I could add is - to an approximation the tide changes 6m in 6hrs - so that’s 0.25m in 15mins - leave at that as you are searching around the “turning-points” - slow changes - add a bit for real effects of wind changing, etc. - so could throw-out data-points where more than 0.4m change in 15mins Reason I put “broadly it’s there”. Must inspect data for improbable-and-isolated data points. Then arrive at what I am going to plot and declare that, to the best of ability, according to declared criteria “this is it”. I have once heard “… and we can see what it is that you have computed”. Massively tipped things in my favour, leaving a feeling of certainty about what they were looking at. These tools would surely be very useful to look for trends which if seen you investigate. Mindful of “Texas sharpshooting” ! :-) Thanks, Rich Smith > On 24 Apr 2023, at 13:56, Tom via list <list@xxxxxxxxxxxxx> wrote: > > Perhaps moving up to AI might help. Theres a free app called Weka which comes with > free online courses and also an app called Moa which can eat time line data and > process it as it arrives. I did a course a few years ago and remember it had > mechanisms for handling date with time based cycles and all sorts of goodies for > finding and marking outliers and removing them from datasets.. > I think you'll find high tides will break records regularly from now on - low > tides breaking records probably means a tsunami is on the way! > Cheers cheerily > Tom te tom te tom > On 24/04/2023 07:41, rds_met wrote: >> Hello all >> >> I had a bit of complete relaxation time yesterday and thoughts >> flowed... >> >> I saw you could do the lot with "awk" - the unix toolkit awk command / >> utility. >> >> "Rough" first output. >> http://weldsmith.co.uk/temp/tide_n22_hi_lo_all.png >> >> Broadly "it's there"; however there are problems where the tide gauge >> was mulfunctioning (a tide making 11.something m above Lowest >> Astronomical Tide would if it really happened make the news!!!). >> Etc. >> >> The horizontal axis is the tide-gauge's sequential 15-minute >> data-record increments in a year >> (* 365 24 4) ;; 35040 >> >> I would help a more sophisticated "data-crawling" program if there >> were inclination to make one. >> >> Regards, >> Rich Smith >> >> -- >> The Mailing List for the Devon & Cornwall LUG >> FAQ: https://www.dcglug.org.uk/faq/ > > > -- > The Mailing List for the Devon & Cornwall LUG > FAQ: https://www.dcglug.org.uk/faq/ -- The Mailing List for the Devon & Cornwall LUG FAQ: https://www.dcglug.org.uk/faq/