Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Updated b3gbi #2

Merged
merged 8 commits into from
Sep 11, 2024
Merged

Updated b3gbi #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
a306955
new version
wlangera Jul 22, 2024
5d88ca4
create cube
wlangera Jul 22, 2024
2c6d465
add multispecies example
wlangera Jul 22, 2024
337687b
custom dictionary
wlangera Jul 22, 2024
bd5d9cd
update gcube
wlangera Jul 26, 2024
7028d3f
style
wlangera Jul 26, 2024
407449c
update grid type
wlangera Sep 10, 2024
8010abc
update gcube version 0.4.0
wlangera Sep 11, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions inst/en_gb.dic
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,11 @@ Daele
Havenlaan
Langeraert
Teirlinckgebouw
csv
datacube
eea
etc
gbi
gcube
mgrs
reprex
314 changes: 279 additions & 35 deletions source/gcube_integration_for_b3gbi.Rmd
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@ knitr::opts_chunk$set(echo = TRUE)
```

```{r, warning=FALSE, message=FALSE}
# Load packages
## Load packages
library(gcube) # Simulate biodiversity data cubes
library(b3gbi) # Calculate general indicators for biodiversity data cubes

Expand All @@ -27,10 +27,9 @@ library(tidyverse) # Data wrangling and visualisation

# Introduction

We install the [b3gbi](https://github.com/b-cubed-eu/b3gbi/) (version 0.2.1) and [gcube](https://github.com/b-cubed-eu/gcube) (version 0.0.1) packages.
We can simulate biodiversity data cubes with gcube and afterwards calculate general indicators with the b3gbi package.
We want to simulate biodiversity data cubes with [gcube](https://github.com/b-cubed-eu/gcube) and afterwards calculate general indicators with the [b3gbi](https://github.com/b-cubed-eu/b3gbi/) package.

**Why do we want this integration?**
*Why do we want this integration?*

The goal of gcube is to provide a simulation framework for biodiversity data cubes.
Simulation studies offer numerous benefits due to their ability to mimic real-world scenarios in controlled and customizable environments.
Expand All @@ -42,7 +41,11 @@ Varying the different parameters provides insights on the factors influencing fi
b3gbi is an R package that provides functions that calculate general biodiversity indicators from data cubes.
Linking the output of simulated cubes from gcube into the b3gbi workflow is thus an essential step in the investigation of the effects of different data cube parameters on final estimated statistics and trends.

# Input for b3gbi
# b3gbi version 0.2.1

We install the [b3gbi](https://github.com/b-cubed-eu/b3gbi/) (version 0.2.1) and [gcube](https://github.com/b-cubed-eu/gcube) (version 0.0.1) packages.

## Example

The input for the b3gbi package is the location of a CSV file.

Expand All @@ -69,11 +72,10 @@ insect_data
The function `process_cube_old()` seems rather strict regarding the column names.
The new `process_cube()` function is more flexible in this sense.

> For an efficient workflow. These functions should also allow R dataframes as input.
## Try gcube output as input

Let's create a cube with gcube.

```{r}
```r
## Create cube with gcube (4 time points, 1 species)
# Create a polygon to simulate occurrences
polygon <- st_polygon(list(cbind(c(5, 10, 8, 2, 3, 5), c(2, 1, 7, 9, 5, 2))))

Expand All @@ -82,6 +84,7 @@ occurrences_df <- simulate_occurrences(
plgn = polygon,
n_time_points = 4,
seed = 123)
#> [using unconditional Gaussian simulation]

# Detect occurrences
detections_df_raw <- sample_observations(
Expand All @@ -107,10 +110,10 @@ buffered_observations <- st_buffer(

# Define a grid over spatial extend
grid_df <- st_make_grid(
buffered_observations,
square = TRUE,
cellsize = c(1.2, 1.2)
) %>%
buffered_observations,
square = TRUE,
cellsize = c(1.2, 1.2)
) %>%
st_sf() %>%
mutate(intersect = as.vector(st_intersects(geometry, polygon,
sparse = FALSE))) %>%
Expand All @@ -133,22 +136,19 @@ ggplot() +
theme_minimal()
```

We add a taxon name and save the cube (dataframe) as a CSV file so we can try and load id with `b3gbi::process_cube()`.

```{r}
data_path <- here::here("data", "raw")
dir.create(data_path, showWarnings = FALSE, recursive = TRUE)
![](https://i.imgur.com/z7eR9pH.png)<!-- -->

```
## Write out csv
occurrence_cube_df %>%
st_drop_geometry() %>%
mutate(species = "species1",
species_key = "s1") %>%
write_delim(file.path(data_path, "gcube_df.csv"), delim = "\t", na = "")
```
write_delim("gcube_df.csv", delim = "\t", na = "")

```{r}
## Process cube with b3gbi
gcube_data <- process_cube(
cube_name = file.path(data_path, "gcube_df.csv"),
cube_name = "gcube_df.csv",
grid_type = "eea",
force_gridcode = TRUE,
cols_year = "time_point",
Expand All @@ -157,27 +157,271 @@ gcube_data <- process_cube(
cols_scientificName = "species",
cols_minCoordinateUncertaintyInMeters = "min_coord_uncertainty",
cols_speciesKey = "species_key"
)

)
gcube_data
#>
#> Processed data cube for calculating biodiversity indicators
#>
#> Date Range: 1 - 3
#> Single-resolution cube with cell size 10 12 13 14 15 16 19 20 23 26 29 30 36 37 5 7 8 9 1 2 3 4 6 11 17 18 21 22 24 25 27 28 31 32 33 34 35 38 ^2
#> Number of cells: 38
#> Grid reference system: eea
#> Coordinate range:
#> xmin xmax ymin ymax
#> NA NA NA NA
#>
#> Total number of observations: 70
#> Number of species represented: 1
#> Number of families represented: Data not present
#>
#> Kingdoms represented: Data not present
#>
#> First 10 rows of data (use n = to show more):
#>
#> # A tibble: 114 × 9
#> year cellCode obs scientificName minCoordinateUncertaint…¹ taxonKey xcoord
#> <dbl> <chr> <dbl> <dbl> <chr> <chr> <dbl>
#> 1 1 10 1 0.552 species1 s1 NA
#> 2 1 12 3 0.0242 species1 s1 NA
#> 3 1 13 2 0.0217 species1 s1 NA
#> 4 1 14 2 0.233 species1 s1 NA
#> 5 1 15 1 0.584 species1 s1 NA
#> 6 1 16 2 0.0241 species1 s1 NA
#> 7 1 19 1 0.0844 species1 s1 NA
#> 8 1 20 2 0.149 species1 s1 NA
#> 9 1 23 1 0.353 species1 s1 NA
#> 10 1 26 1 0.208 species1 s1 NA
#> # ℹ 104 more rows
#> # ℹ abbreviated name: ¹​minCoordinateUncertaintyInMeters
#> # ℹ 2 more variables: ycoord <dbl>, resolution <chr>


## Calculate an indicator over time
total_occ_ts(gcube_data)
#> Error in if (stringr::str_detect(resolution, "km")) {: the condition has length > 1
```

> Should be able to use a custom grid_type with forced grid code. Now metadata is wrong.
<sup>Created on 2024-07-09 with [reprex v2.1.0](https://reprex.tidyverse.org)</sup>

> This code does not work if you have only 1 time point.
This throws an error by `check_cell_size()`.

We can try and calculate an indicator.
A spatial indicator will not work since we do not have the right spatial metadata.
We try a time series indicator.
**Challenges**

```{r, eval=FALSE}
total_occ_ts(gcube_data)
1. For an efficient workflow `process_cube()` should also allow R dataframes as input
- Currently, gcube dataframes should be stored as CSV files and their paths used as input for `process_cube()`
2. Allow custom grid type in `process_cube()`
- With `force_gridcode = TRUE`
- Now the metadata is incorrect and no indicators can be calculated further upstream
3. Calculate indicators for custom cubes
- I understand visualisation cannot made possible, but at least it should be possible to calculate the indicators since we have the same data type (does not matter if it is eea or mgrs or ..., year or month or time_period or ...)
- Visualisation can be done by the user itself

# b3gbi version 0.2.3

Review for [this pull request](https://github.com/b-cubed-eu/b3gbi/pull/25).
We install the [b3gbi](https://github.com/b-cubed-eu/b3gbi/) (version 0.2.3) and [gcube](https://github.com/b-cubed-eu/gcube) (version 0.4.0) packages.

We create a datacube with **gcube** for 6 species over 6 time points.
First we define the spatial extend.

```{r}
# Create a polygon to simulate occurrences
polygon <- st_polygon(list(cbind(c(500, 1000, 1000, 600, 200, 100, 500),
c(200, 100, 700, 1000, 900, 500, 200))))

# Create grid for grid designation
cube_grid <- st_make_grid(
st_buffer(polygon, 50),
n = c(20, 20),
square = TRUE) %>%
st_sf()

# Visualise
ggplot() +
geom_sf(data = polygon) +
geom_sf(data = cube_grid, alpha = 0) +
theme_minimal()
```

Let's simulate the cube.

```{r}
# Create dataframe with simulation function arguments
multi_species_dataset <- tibble(
species_range = rep(list(polygon), 6),
n_time_points = rep(6, 6),
detection_probability = rep(c(0.8, 0.9, 1), 2),
coords_uncertainty_meters = rep(c(25, 30, 50), 2),
grid = rep(list(cube_grid), 6),
seed = 123
)

# Add taxonomic hierarchy and generate cube
map_occ_cube_df <- multi_species_dataset %>%
generate_taxonomy(num_genera = 4, num_families = 2, seed = 123) %>%
map_simulate_occurrences() %>%
map_sample_observations() %>%
map_filter_observations() %>%
map_add_coordinate_uncertainty() %>%
map_grid_designation(nested = FALSE) %>%
select(-all_of(names(multi_species_dataset))) %>%
select(-occurrences, -observations_total, -observations)

glimpse(map_occ_cube_df)
```

```{r, echo=FALSE}
try(expr = total_occ_ts(gcube_data))
This time we do not write out a csv, but we use the dataframe to process the cube.

```{r}
# Process cube with b3gbi
gcube_data <- process_cube(
cube_name = map_occ_cube_df,
grid_type = "custom",
cols_year = "time_point",
cols_cellCode = "cell_code",
cols_occurrences = "n",
cols_scientificName = "species",
cols_minCoordinateUncertaintyInMeters = "min_coord_uncertainty",
cols_kingdom = "kingdom",
cols_family = "family",
cols_speciesKey = "species_key"
)
gcube_data
```

This throws an error by `check_cell_size()`.
```{r}
total_occ_ts(gcube_data)
```

reprex:

```r
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
library(sf)
#> Linking to GEOS 3.12.1, GDAL 3.8.4, PROJ 9.3.1; sf_use_s2() is TRUE
library(gcube)
library(b3gbi)

# Create a polygon to simulate occurrences
polygon <- st_polygon(list(cbind(c(500, 1000, 1000, 600, 200, 100, 500),
c(200, 100, 700, 1000, 900, 500, 200))))

# Create grid for grid designation
cube_grid <- st_make_grid(
st_buffer(polygon, 50),
n = c(20, 20),
square = TRUE) %>%
st_sf()

# Create dataframe with simulation function arguments
multi_species_dataset <- tibble(
plgn = rep(list(polygon), 6),
n_time_points = rep(6, 6),
detection_probability = rep(c(0.8, 0.9, 1), 2),
coords_uncertainty_meters = rep(c(25, 30, 50), 2),
grid = rep(list(cube_grid), 6),
seed = 123
)

# Add taxonomic hierarchy and generate cube
map_occ_cube_df <- multi_species_dataset %>%
generate_taxonomy(num_genera = 4, num_families = 2, seed = 123) %>%
map_simulate_occurrences() %>%
map_sample_observations() %>%
map_filter_observations() %>%
map_add_coordinate_uncertainty() %>%
map_grid_designation(nested = FALSE) %>%
select(-all_of(names(multi_species_dataset))) %>%
select(-occurrences, -observations_total, -observations)
#> [1] [using unconditional Gaussian simulation]
#> [2] [using unconditional Gaussian simulation]
#> [3] [using unconditional Gaussian simulation]
#> [4] [using unconditional Gaussian simulation]
#> [5] [using unconditional Gaussian simulation]
#> [6] [using unconditional Gaussian simulation]

# Process cube with b3gbi
gcube_data <- process_cube(
cube_name = map_occ_cube_df,
grid_type = "none",
cols_year = "time_point",
cols_cellCode = "id",
cols_occurrences = "n",
cols_scientificName = "species",
cols_minCoordinateUncertaintyInMeters = "min_coord_uncertainty",
cols_kingdom = "kingdom",
cols_family = "family",
cols_speciesKey = "species_key"

)
gcube_data
#>
#> Simulated data cube for calculating biodiversity indicators
#>
#> Date Range: 1 - 5
#> Number of cells:
#> Grid reference system: none
#> Coordinate range:
#> NULL
#>
#> Total number of observations: 1382
#> Number of species represented: 6
#> Number of families represented:
#>
#> Kingdoms represented:
#>
#> First 10 rows of data (use n = to show more):
#>
#> # A tibble: 12,000 × 13
#> scientificName taxonKey genus family order class phylum kingdom year id
#> <chr> <dbl> <chr> <chr> <chr> <chr> <chr> <chr> <dbl> <chr>
#> 1 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 106
#> 2 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 109
#> 3 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 113
#> 4 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 117
#> 5 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 119
#> 6 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 124
#> 7 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 131
#> 8 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 134
#> 9 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 147
#> 10 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 154
#> # ℹ 11,990 more rows
#> # ℹ 3 more variables: obs <dbl>, minCoordinateUncertaintyInMeters <dbl>,
#> # geometry <POLYGON>

# Try calculate time series indicator
total_occ_ts(gcube_data)
#> Biodiversity indicator time series
#>
#> Name of indicator: Total Occurrences
#>
#> Date Range: 1 - 5
#>
#> Coordinate range represented:
#> xmin xmax ymin ymax
#> "NA" "NA" "NA" "NA"
#>
#> Number of species represented: 6
#> Kingdoms represented: NA
#>
#> First 10 rows of data (use n = to show more):
#>
#> # A tibble: 5 × 2
#> year diversity_val
#> <dbl> <dbl>
#> 1 1 244
#> 2 2 326
#> 3 3 198
#> 4 4 282
#> 5 5 332
```

> Why do we need to know the spatial resolution for a temporal indicator?
<sup>Created on 2024-07-26 with [reprex v2.1.0](https://reprex.tidyverse.org)</sup>