bgautijonsson / bggjphd

An R package to go along with my PhD research

About this page

Information about this package and why it was created can be found at my website

Installing the package

The package can be installed with pak or remotes. I recommend using pak, since it can also install any non-R dependencies that might be required. I’ve started to move all my package installation to pak, and use it for example to set up my Docker container that sets up an R environment along with any needed dependencies for bggjphd.

# using devtools
install.packages("remotes")
remotes::install_github("bgautijonsson/bggjphd")

# using pak
install.packages("pak")
pak::pak("bgautijonsson/bggjphd")

Data

The bggjphd package comes with some dataset

library(bggjphd)
#> 
#> Attaching package: 'bggjphd'
#> The following object is masked from 'package:datasets':
#> 
#>     precip

precip

The precip dataset contains maximum hourly precipitation calculated yearly for each of the locations in the UKCP Local Projections on a 5km grid over the UK for 1980-2080 from the CEDA archive.

skimr::skim(precip)

Name	precip
Number of rows	2723040
Number of columns	3
_______________________
Column type frequency:
numeric	3
________________________
Group variables	None

Data summary

Variable type: numeric

skim_variable	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
year	0	1	2030.81	32.87	1981.00	1996.00	2030.50	2065.00	2080.00	▇▁▇▁▇
station	0	1	21960.50	12678.61	1.00	10980.75	21960.50	32940.25	43920.00	▇▇▇▇▇
precip	0	1	12.88	5.90	0.79	9.05	11.49	15.14	107.88	▇▁▁▁▁

stations

The stations dataset contains information about the stations at which the observations from precip are recorded.

skimr::skim(stations)

Name	stations
Number of rows	43920
Number of columns	5
_______________________
Column type frequency:
numeric	5
________________________
Group variables	None

Data summary

Variable type: numeric

skim_variable	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
station	0	1	21960.50	12678.76	1.00	10980.75	21960.50	32940.25	43920.00	▇▇▇▇▇
proj_x	0	1	90.50	51.96	1.00	45.75	90.50	135.25	180.00	▇▇▇▇▇
proj_y	0	1	122.50	70.44	1.00	61.75	122.50	183.25	244.00	▇▇▇▇▇
latitude	0	1	54.98	3.16	49.31	52.25	54.98	57.71	60.53	▇▇▇▇▇
longitude	0	1	-4.36	4.10	-12.77	-7.86	-4.35	-0.82	3.41	▅▇▇▇▆

station_estimates

The station_estimates dataset contains parameter estimates and Hessian matrices from generalized maximum likelihood estimation of GEV models with trend at each of the locations. The results are stored in list-columns which need to be unnested for analysis. Inside the list columns the results are stored in long format.

station_estimates |> 
  dplyr::select(par) |> 
  tidyr::unnest(par) |> 
  dplyr::group_by(name) |> 
  skimr::skim()

Name	dplyr::group_by(…)
Number of rows	175680
Number of columns	2
_______________________
Column type frequency:
numeric	1
________________________
Group variables	name

Data summary

Variable type: numeric

skim_variable	name	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
value	psi	0	1	2.26	0.15	1.80	2.14	2.27	2.38	2.82	▁▆▇▃▁
value	tau	0	1	-0.99	0.13	-1.63	-1.08	-0.99	-0.90	-0.52	▁▁▇▇▁
value	phi	0	1	0.03	0.09	-0.47	-0.02	0.04	0.09	0.37	▁▁▇▇▁
value	gamma	0	1	0.00	0.00	0.00	0.00	0.00	0.00	0.01	▁▅▇▂▁

station_estimates |> 
  dplyr::select(hess) |> 
  tidyr::unnest(hess) |> 
  dplyr::group_by(name1, name2) |> 
  skimr::skim()

Name	dplyr::group_by(…)
Number of rows	702720
Number of columns	3
_______________________
Column type frequency:
numeric	1
________________________
Group variables	name1, name2

Data summary

Variable type: numeric

skim_variable	name1	name2	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
value	psi	psi	0	1	525.61	158.38	164.76	411.31	500.66	609.24	2055.00	▇▅▁▁▁
value	psi	tau	0	1	23.75	44.19	-235.65	-1.95	24.97	49.89	304.88	▁▂▇▁▁
value	psi	phi	0	1	92.75	19.80	38.90	78.71	90.85	104.64	272.54	▆▇▁▁▁
value	psi	gamma	0	1	20771.41	7006.04	6549.91	15911.09	19460.66	24063.30	87420.99	▇▃▁▁▁
value	tau	psi	0	1	23.75	44.19	-235.65	-1.95	24.97	49.89	304.88	▁▂▇▁▁
value	tau	tau	0	1	117.66	7.28	91.66	112.69	117.21	122.03	173.72	▁▇▂▁▁
value	tau	phi	0	1	3.98	16.21	-89.96	-5.87	5.20	15.28	79.58	▁▁▇▅▁
value	tau	gamma	0	1	-4289.97	2078.75	-19160.63	-5464.97	-4146.20	-2951.75	6433.63	▁▁▇▆▁
value	phi	psi	0	1	92.75	19.80	38.90	78.71	90.85	104.64	272.54	▆▇▁▁▁
value	phi	tau	0	1	3.98	16.21	-89.96	-5.87	5.20	15.28	79.58	▁▁▇▅▁
value	phi	phi	0	1	221.71	63.53	59.76	176.47	210.82	254.49	705.14	▅▇▁▁▁
value	phi	gamma	0	1	5605.60	1820.66	541.58	4276.29	5412.46	6704.33	17141.02	▂▇▂▁▁
value	gamma	psi	0	1	20771.41	7006.04	6549.91	15911.09	19460.66	24063.30	87420.99	▇▃▁▁▁
value	gamma	tau	0	1	-4289.97	2078.75	-19160.63	-5464.97	-4146.20	-2951.75	6433.63	▁▁▇▆▁
value	gamma	phi	0	1	5605.60	1820.66	541.58	4276.29	5412.46	6704.33	17141.02	▂▇▂▁▁
value	gamma	gamma	0	1	1471151.77	532303.18	349192.83	1115369.11	1377977.19	1713948.25	7597034.07	▇▂▁▁▁

other data

There are other objects stored as data in bggjphd, but these are mostly kept there for use with the ms_smooth() function, and are binary versions of sparse Matrix:: matrices.