Skip to content

Commit

Permalink
docs: editing and refining datacheck visualizations for NP step 2
Browse files Browse the repository at this point in the history
  • Loading branch information
@annaramji
annaramji committed Aug 19, 2024
1 parent 7f8e058 commit 7a0d66c
Showing 1 changed file with 105 additions and 122 deletions.
227 changes: 105 additions & 122 deletions globalprep/np/v2024/STEP2_np_weighting_prep.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -336,7 +336,12 @@ readr::write_csv(prod_weights, here(current_np_dir, "output", "np_product_weight


### Datacheck

Note: we're not comparing data for a specific year, e.g., 2019, as we do in many data checks. The data processing that creates `np_product_weights.csv` fills the `year` column with the scenario year (e.g., 2024), which means that the previous year's `np_product_weights.csv` and the current year's weights cannot be compared by filtering to a specific year.


```{r fig.height=5, fig.width=8}
# Read in previous and current years' np_product_weights to compare
#old_prod_weights <- read_csv(file.path(prep, paste0("../v", prep_year-3, "/output/np_product_weights.csv")))
#old_prod_weights <- read_csv(file.path(prep, paste0("../v", prep_year-2, "/output/np_product_weights.csv")))
Expand All @@ -352,20 +357,20 @@ check <- prod_weights %>%
# note -- these are across all years...
plot(check$new_weight, check$weight, main = "All Products")
abline(0,1, col="red")
abline(0,1, col = "red")
check_sw <- check %>%
filter(product == "seaweeds")
plot(check_sw$new_weight, check_sw$weight, main = "Seaweeds")
abline(0,1, col="red")
abline(0,1, col = "red")
check_orn <- check %>%
filter(product == "ornamentals") %>%
mutate(difference = new_weight - weight)
plot(check_orn$new_weight, check_orn$weight, main = "Ornamentals")
abline(0,1, col="red")
abline(0,1, col = "red")
check_fofm <- check %>%
Expand All @@ -374,9 +379,8 @@ check_fofm <- check %>%
plot(check_fofm$new_weight, check_fofm$weight, main = "Fish Oil / Fish Meal (FOFM)")
abline(0, 1, col = "red") ## these change because of new SAUP fisheries catch data...
max(check$diff, na.rm = TRUE)
min(check$diff, na.rm = TRUE)
# max(check$diff, na.rm = TRUE)
# min(check$diff, na.rm = TRUE)
top_10_diffs <- check %>% arrange(desc(abs(diff))) %>% head(n = 10)
top_10_diffs %>% relocate(diff, .after = "rgn_id")
Expand Down Expand Up @@ -407,15 +411,14 @@ test2 <- harvest_tonnes_usd_old %>%
# test2
# v2024: Djibouti (rgn 46) does increase in production
gf_orn <- read_csv(here(current_np_dir, "output", "np_ornamentals_harvest_tonnes_gf.csv"))
#gf_orn <- read_csv(here(current_np_dir, "output", "np_ornamentals_harvest_tonnes_gf.csv"))
```





Old old check:

```{r}
Expand All @@ -437,61 +440,6 @@ plot(check_older$weight, check_older$older_weight, main = "v2023 vs v2022"); abl
plot(check$new_weight, check$weight, main = "v2024 vs v2023"); abline(0,1, col="red")
# interactive plot:
weight_comp_plot <- ggplot(check_older, aes(x = older_weight, y = weight,
color = product,
text = paste("Region ID:", rgn_id) #, "<br>Product:", product)
#fill = rgn_id
)) +
geom_abline(intercept = 0, slope = 1, color = "#3498DB", linewidth = 1.3) + # Change color and size of line
geom_point(size = 2) + # Increase point size
scale_color_manual(values = c("fish_oil" = "#edae49", "seaweeds" = "#386641", "ornamentals" = "#83c5be")) +
theme_minimal() +
labs(x = "v2022 Weights", y = "v2023 Weights",
title = "Comparing NP Product Weights",
subtitle = "") +
theme(legend.position = "none")
plotly::ggplotly(weight_comp_plot) %>% plotly::layout(yaxis = list(hoverformat = '.4f'), xaxis = list(hoverformat = '.4f'))
weight_comp_plot_new <- ggplot(check, aes(x = weight, y = new_weight,
fill = product,
text = paste("Region ID:", rgn_id, "<br>Product:", product)
#fill = rgn_id
)) +
geom_abline(intercept = 0, slope = 1, color = "#3498DB", linewidth = 1.3) + # Change color and size of line
geom_point(size = 2) + # Increase point size
scale_fill_manual(values = c("fish_oil" = "#edae49", "seaweeds" = "#386641", "ornamentals" = "#83c5be")) +
theme_minimal() +
labs(x = "v2022 Weights", y = "v2023 Weights",
title = "Comparing NP Product Weights",
subtitle = "") +
theme(legend.position = "none")
p <- plotly::ggplotly(weight_comp_plot, tooltip = "text") %>%
plotly::layout(
hovermode = "closest",
yaxis = list(hoverformat = '.4f'), xaxis = list(hoverformat = '.4f'))
# Custom hover template
p$x$data <- lapply(p$x$data, function(trace) {
if (trace$mode == "markers") {
trace$hovertemplate <- paste(
"%{text}<br>",
"<b>Product</b>: %{fill}<br>",
"<b>Old Weight</b>: %{x:.4f}<br>",
"<b>New Weight</b>: %{y:.4f}<br>"
)
}
return(trace)
})
p
```


Expand All @@ -500,26 +448,25 @@ p
library(plotly)
library(RColorBrewer)
color_scale <- c("fish_oils" = "#edae49", "seaweeds" = "#00798c", "ornamentals" = "#d1495b")
test_comp <- check_older %>% plotly::plot_ly(x = ~older_weight, y = ~weight,
type = "scatter", mode = "markers") %>%
add_trace(
# text = ~paste("Region ID:", rgn_id, "<br>Product:", product),
# hoverinfo = "text+x+y",
text = (~rgn_id, ~product),
marker = list(
color = ~product,
colors = color_scale
),
hovertemplate = paste(
'<b>Region ID</b>: %{text}',
'<br><b>Old Weight</b>: %{x:.2f}',
'<br><b>New Weight</b>: %{y:.2f}<br>'
# text = ~paste("Region ID:", rgn_id #, "<br>Product:", product
# ),
color = ~rgn_id,
# marker = list(
# color = ~product,
# ),
hoverinfo = "text+x+y",
hovertemplate = paste(
'<b>Region ID</b>: %{color}',
'<br><b>Old Weight</b>: %{x:.2f}',
'<br><b>New Weight</b>: %{y:.2f}<br>'
),
# showlegend = F
) %>%
# add AB line
add_trace(
x = c(0,1),
y = c(0,1),
Expand All @@ -539,42 +486,76 @@ test_comp
- RAM lost some stocks, so some stocks get removed in v4.65 (2024 data download); also got some new stocks



### New datacheck: interactive plot

```{r}
# ======== Setup ==========================================================
# (mostly copied from earlier setup with addition of library(plotly))
# Read in packages necessary for visualization (in case you cleared your environment)
library(ggplot2)
library(plotly)
library(dplyr)
library(tidyr)
library(here)
color_scale <- c("fish_oils" = "#edae49", "seaweeds" = "#00798c", "ornamentals" = "#d1495b")
# Set scenario year, reproducible file paths
scen_year_number <- 2024 # update this!!
scenario_year <- as.character(scen_year_number)
v_scen_year <- paste0("v", scenario_year)
current_np_dir <- here::here("globalprep", "np", v_scen_year)
previous_np_dir <- here::here("globalprep", "np", paste0("v", scen_year_number - 1))
# ----- Read in NP weights data -----
# Read in current, previous, and 2 years previous data
old_prod_weights <- read_csv(here(previous_np_dir, "output", "np_product_weights.csv"))
prod_weights <- read_csv(here(current_np_dir, "output", "np_product_weights.csv"))
check <- prod_weights %>%
rename("new_weight" = "weight") %>%
left_join(old_prod_weights, by = c("rgn_id", "product")) %>%
mutate(diff = new_weight - weight) %>%
left_join(rgns_eez)
#old_prod_weights <- read_csv(here(previous_np_dir, "output", "np_product_weights.csv"))
older_prod_weights <- read_csv(here("globalprep", "np", paste0("v", scen_year_number - 2), "output", "np_product_weights.csv"))
check_older <- older_prod_weights %>%
rename("older_weight" = "weight") %>%
left_join(old_prod_weights, by = c("rgn_id", "product")) %>%
mutate(diff = weight - older_weight) %>%
left_join(rgns_eez)
# ======== Plot new NP weights data vs. previous year's data ==================
# Assuming check_older is your dataframe
# If not, replace check_older with your actual dataframe name
# Create the ggplot object
g <- ggplot(check_older, aes(x = older_weight, y = weight, color = product,
# ---- Create ggplot object with `text` aes argument to prep for plotly ----- #
g_new <- ggplot(check, aes(x = weight, y = new_weight, color = product,
text = paste("<b>Region ID:</b>", rgn_id, "<br><b>Product:</b>", product #,
#"<br>Old weight:", older_weight, "<br>New weight:", weight
))) +
geom_point() +
geom_abline(intercept = 0, slope = 1, color = "red", linetype = "dashed") +
geom_point(alpha = 0.8) +
geom_abline(intercept = 0, slope = 1, alpha = 0.7,
color = "red", linetype = "dashed") +
scale_color_manual(values = c("fish_oil" = "#edae49", "seaweeds" = "#386641", "ornamentals" = "#83c5be")) +
labs(title = "v2023 vs v2022 Weights by Product",
x = "v2022 Weight",
y = "v2023 Weight",
labs(title = "v2024 vs v2023 Natural Product Weights by Product",
x = "v2023 Weight",
y = "v2024 Weight",
color = "Product") +
theme_minimal()
# Convert to plotly
p <- ggplotly(g, tooltip = "text") %>%
# ---- Convert to plotly ------------------------------ #
p_new <- ggplotly(g_new, tooltip = "text") %>%
layout(
hovermode = "closest",
xaxis = list(hoverformat = ".4f"),
yaxis = list(hoverformat = ".4f")
xaxis = list(hoverformat = ".4f"), # customize number formatting
yaxis = list(hoverformat = ".4f") # (number of places after decimal)
)
# Custom hover template
p$x$data <- lapply(p$x$data, function(trace) {
# ------- Create custom hover template ---------------- #
p_new$x$data <- lapply(p_new$x$data, function(trace) {
if (trace$mode == "markers") {
trace$hovertemplate <- paste(
"%{text}<br>",
Expand All @@ -585,47 +566,48 @@ p$x$data <- lapply(p$x$data, function(trace) {
return(trace)
})
# Display the plot
p
```
# Display customized plot
p_new
# ========= Plot old vs. older NP weights data ================================
# Plot previous year's NP weights data vs. year before previous year's data
```{r}
library(ggplot2)
library(plotly)
library(dplyr)
color_scale <- c("fish_oils" = "#edae49", "seaweeds" = "#00798c", "ornamentals" = "#d1495b")
# Assuming check_older is your dataframe
# If not, replace check_older with your actual dataframe name
#color_scale <- c("fish_oils" = "#edae49", "seaweeds" = "#00798c", "ornamentals" = "#d1495b")
# Create the ggplot object
g <- ggplot(check, aes(x = weight, y = new_weight, color = product,
text = paste("<b>Region ID:</b>", rgn_id, "<br><b>Product:</b>", product #,
#"<br>Old weight:", older_weight, "<br>New weight:", weight
))) +
geom_point() +
geom_abline(intercept = 0, slope = 1, color = "red", linetype = "dashed") +
# ---- Create ggplot object with added `text` field to prep for plotly ------ #
g_old <- ggplot(check_older, aes(
x = older_weight, y = weight, color = product,
text = paste("<b>Region ID:</b>", rgn_id, "<br><b>Product:</b>", product #,
#"<br>Old weight:", older_weight, "<br>New weight:", weight
)
)) +
# add scatterplot points
geom_point(alpha = 0.8) +
# add ab line
geom_abline(intercept = 0, slope = 1, alpha = 0.7,
color = "red", linetype = "solid") +
# customize colors
scale_color_manual(values = c("fish_oil" = "#edae49", "seaweeds" = "#386641", "ornamentals" = "#83c5be")) +
labs(title = "v2024 vs v2023 Natural Product Weights by Product for 2019",
x = "v2023 2019 Weight",
y = "v2024 2019 Weight",
# update labels
labs(title = "v2023 vs v2022 Natural Products Weights by Product",
x = "v2022 Weights",
y = "v2023 Weights",
color = "Product") +
# set base theme
theme_minimal()
# Convert to plotly
p <- ggplotly(g, tooltip = "text") %>%
# ---- Convert to plotly ------------------------------ #
p_old <- ggplotly(g_old, tooltip = "text") %>%
layout(
hovermode = "closest",
xaxis = list(hoverformat = ".4f"),
xaxis = list(hoverformat = ".4f"), # customize number formatting
yaxis = list(hoverformat = ".4f")
)
# Custom hover template
p$x$data <- lapply(p$x$data, function(trace) {
# ------- Custom hover template ----------------------- #
p_old$x$data <- lapply(p_old$x$data, function(trace) {
if (trace$mode == "markers") {
trace$hovertemplate <- paste(
"%{text}<br>",
Expand All @@ -636,7 +618,8 @@ p$x$data <- lapply(p$x$data, function(trace) {
return(trace)
})
# Display the plot
p
# Display customized plot!
p_old
```


0 comments on commit 7a0d66c

Please sign in to comment.