search

mcolvin / Bluff-Lake-Project

Bluff Lake Project Repository
0 stars 0 forks source link

stupid dates #22

Closed mcolvin closed 4 years ago

mcolvin commented 4 years ago

@VictoriaStarnes

  1. Confirm dates are right for loggers that are read into R
  2. missing values for intake?

image

Missing 10:00 model_data[ year==2020&doy==26&hour%in%c(8,9,10),] Missing 10:30 model_data[ year==2020&doy==40&hour%in%c(9,10,11),] Missing 1:30 to 3:00 model_data[ year==2020&doy==68 &hour%in%c(0,1,2,3,4),]

image

VictoriaStarnes commented 4 years ago
  1. Wrong origin start date in original date conversion.
  2. Subset sections using as.Date() do not include time. Need to use as.POSIXct() in subset to pull date and time.
  3. NaN's are at the beginning/end of logger profiles. Sometimes the first or last few logger recordings in a raw file only have a time recording, so I just delete these rows. I used na.approx() to fill in gaps. (15 values)
mcolvin commented 4 years ago

na.approx is throwing an error.

VictoriaStarnes commented 4 years ago

I FIXED IT!

The issue with na.approx was that it does not like trailing or leading NA values. Once those were removed that function ran fine. I then tried the model and it still wouldn't run. We had a gap on March 8th. After trying FOREVER to fill in that gap with NA values I realized that it was caused by daylight savings time. I then adjusted our continuous time to ignore springing forward. Then I was FINALLY able to run the model! It looks quite a bit better after the adjusted time and shifted start date.

image