Citizen Scientists

• The Bureau of Meteorology has been getting reports about weird temperatures in Iowa
• Apparently, there is a strange spatial pattern in the data
• We get some citizen scientists to measure data at their home and report back

NOT SO FAST CHUMP

The citizen scientist data

toy_temp

990 citizen scientists participated

Visualisation goals

• We need to look at the data and identify:
  • The spatial trend (does it exist or not)
  • The statistical strength of the spatial trend

Approach needs to work for…

…and…

Uncertainty visualisation should…

Can also think of it as…

We could just plot the data…

Benefits of plotting the original data

• Plots of the original data can reveal it’s limitations
  • Can highlight small sample sizes, high variability, missing values, etc.
• Understanding these limitations can prevent false conclusions

But the original data isn’t always…

• Available
  • e.g. anonymised data, theoretical values, etc.
• Relevant
  • Might need to use a sampling distribution rather than original data
• Deterministic
  • e.g. bounded data, estimated values, etc.

Going back to the example

After we send in our data plot The Bureau of Meteorology call us and let us know that we are only looking for a trend in the average values of the counties

• Need to use the sampling distribution
  • We do not have data for the sampling distribution
  • Usually we do not bother, instead we do something like…

Estimate the county mean

# Calculate County Mean
toy_temp |> 
  group_by(county_name) |>
  summarise(temp_mean = mean(recorded_temp)) 

Visualise with a choropleth map

But what if the error is worse?

Spot the difference

Solution: add an axis for uncertainty

Solution: blend the colours together!

Solution: simulate a sample

Storing uncertain data

Thank god for distributional

toy_temp_est <- toy_temp |> 
  group_by(county_name) |>
  summarise(temp_dist = dist_normal(mu = mean(recorded_temp),
                                    sigma = sd(recorded_temp)/sqrt(n())))

What does it do?

• distributional lets you store distributions in a tibble as distribution objects

class(toy_temp_est$temp_dist)

[1] "distribution" "vctrs_vctr"   "list"        

• Can also mix different distributions together (don’t all need to be the same family)
• Can use a sample rather than a theoretical distribution
  • e.g. bootstrapping

Code to make pixel map

Ideal pixelation map code

• the ggplot recognises the random variable input, and changes the visualisation accordingly
• Again, touch as few ggplot settings as possible

# Psudo Code
ggplot(data) |>
  geom_sf(aes(geometry = geometry,
              fill = random_variable)) 

Vizumap code

# load the package
library(Vizumap)
library(sf)
sf_use_s2(FALSE)

# Step 1: Format data using bespoke data formatting function
data <- read.uv(data = original_data, 
                estimate = "mean", 
                error = "standard_error")

# Step 2: Pixelate the shapefile
pixelation <- pixelate(geoData = geometry_data, 
                       id = "ID", 
                       # improved - set number of pixels
                       pixelSize = 100)


# Step 3: Build pixel map
pixel_map <- build_pmap(data = data, 
                         distribution = "normal", 
                         pixelGeo = pixelation, 
                         id = "ID", 
                         # You can only use a set palette
                         palette = "Oranges"
                         border = geometry_data)

# Step 4: Print pixel map
view(pixel_map)

Can we plot the pixel map now?

• Have tidy format for distributions
• Now need to extend ggplot2 to implement it

ggplot2 uses the grammar of graphics

It is designed to take in data

Not theoretical distributions

This is what ggdibbler is for

The ggdibbler package

ggdibbler Example

library(ggdibbler)
toy_temp_dist |> 
  ggplot() + 
  geom_sf_sample(aes(geometry = county_geometry,
                     fill=temp_dist))

Wow, look at that software go

ggplot(toy_temp_dist) +
  geom_sf_sample(aes(geometry=county_geometry, fill=temp_dist),  linewidth=0, n=7) +
  geom_sf(aes(geometry = county_geometry), fill=NA, linewidth=0.5, colour="white") +
  theme_minimal() +
  scale_fill_distiller(palette = "YlOrRd", direction= 1) +
  xlab("Longitude") +
  ylab("Latitude") +
  labs(fill = "Temperature") +
  ggtitle("A super cool and customised plot")

Wow, look at that software go x2

ggplot(toy_temp_dist) +
  geom_sf_sample(aes(geometry=county_geometry, fill=temp_dist),  linewidth=0, n=7) +
  geom_sf(aes(geometry = county_geometry), fill=NA, linewidth=0.5, colour="white") +
  theme_minimal() +
  scale_fill_distiller(palette = "YlOrRd", direction= 1) +
  xlab("Longitude") +
  ylab("Latitude") +
  labs(fill = "Temperature") +
  ggtitle("A super cool and customised plot")

If you care about software

If you care about theory