arc2 data processing

[joseepoirier]

There are currently 2 files for your review:

1) to pull the .nc and extract the values (_data_pull.R) 2) to compute various variables including season onset/cessation/duration. (_data_analysis.R)

There is a .Rmd (coming to a PR near you in the near future) into which these two files will be sourced, and in which the reporting metrics will be computed per month etc. and plotted.

Notes:

• 64 cells display no data. These are the touching cells that weren't included in the .nc yet. 279 cells have their centroids within the Southern region and thus are included. In total, there are 343 cells of interest.
• The current dataset only goes until 20 March so no cessation date is available yet.

In files # 1 and # 2, please poke around more intensely around these points:

• date "translations": are the numerical start_date and end_date correct?
• alignment in rollsums (eg: since the definition of onset requires it be after 1 Nov, I used align=left with a lead 1 so the sum would reflect the next 10 days excluding today)
• onset and cessation dates computations
• in-season inclusions: because of the missing data all in_seas_bin values are "unknown". While for most cells we'll eventually be able to determine their actual onset_date and cessation_date and whether a given datapoint was during the rainy season, I think that there is a logical possibility that a cell may not meet the strict definition at all in a given year. Not sure if we should plan for the eventuality or if it's so unlikely (and we confirm it doesn't happen once we have the data) that we can ignore it for now. I planned for the eventuality by assigning the cells arbitrary dates to keep the calculation simple (instead of average of close cells' dates) and still include them in the weekly and monthly totals. Not sure this is desirable in the final output, figured this would give us something concrete to talk through our options.

Thanks!

[joseepoirier]

Thanks so much @caldwellst ! Glad to have had another set of eyes on the accuracy of the calculations. Bonus for pre-reviewing improvements to the code itself, too. :)

Sorry I didn't mean to have you review the .Rmd's yet..

[caldwellst]

No worries at all! I will double check all calculations again in the next review, because my code eyes are just dead now.

[joseepoirier]

No worries at all! I will double check all calculations again in the next review, because my code eyes are just dead now.

Completely understood! If it makes sense to wait until we have the full report ready for review and/or to pull a more complete dataset, don't hesitate to wait. At least we have a first pass so I can keep building on it.

[caldwellst]

Hey Josee! Got pinged again on this, but not sure if ready for review. Won't go through it yet, but let me know if you want it reviewed or any specific part reviewed and I will do so!

[joseepoirier]

@caldwellst Indeed it is not ready for your review. Just resumed work on it. Will ping you directly when it is ready.

[joseepoirier]

@caldwellst I am going to close this PR now. The full report will soon be ready for review and I'll send you a separate PR for that to include the Rmd.