Echos in The Night: California’s Hidden Neighbors

Mexican free-tailed bat or Brazilian free-tailed bat (Tadarida brasiliensis)

Introduction:

Few things ruin a warm evening faster than the whine of a mosquito. As California’s climate shifts, mosquito season has been starting earlier, fueled by warmer and wetter conditions. Fortunately, an unexpected group of nighttime heroes is already on the job: bats. These nocturnal neighbors are powerful pest-control allies, consuming enormous numbers of insects each night, often eating 25–50% of their body mass before sunrise.¹ In doing so, bats save the agricultural industry billions of dollars each year in reduced pesticide use and crop damage.

Insect-eating bats make up the majority of California’s bat species, and their ecological contributions extend far beyond pest control. Some bats pollinate economically important plants, while others serve as indicators of ecosystem health.²³ For example, species like the northern hoary bat roost exclusively in tree foliage, making their presence a sign of high-quality habitat. Yet despite these important ecosystem services, bats remain largely invisible to most people.

This is where citizen science becomes invaluable. Over the past 27 years, 975 observers on iNaturalist have documented bat sightings across California. These observations help reveal where bats are being seen, which species are present, and when sightings peak throughout the year. Using this community-generated data, my visualization explores the hidden patterns of bat observations across the state and highlights the role these often-overlooked animals play in California’s ecosystems.

To understand what citizen scientists have revealed about California’s bats, I analyzed 1,843 observations collected between 1998 and 2025. This informed the design of my infographic, and three patterns emerged that paint a picture of remarkable diversity coexisting with human communities across the state.

What species are out there?

The data reveal at least 993 research-grade confirmed bat species documented by community observers, with the Mexican Free-tailed Bat leading observations. What’s particularly compelling is that 36% of observations are still awaiting species-level identification, labeled as “Unidentified Mouse-eared Bat” or “Unidentified Evening Bat.” Rather than being a data limitation, this tells an important story about citizen science in action: everyday people are actively documenting bat encounters even without expert knowledge, contributing valuable presence data that experts can later verify and refine.

Where are they found?

The heatmap and spatial pattern map reveals bats aren’t confined to remote wilderness, they’re distributed across California’s diverse landscapes, from coastal edges to inland valleys. Observation density clusters around populated areas, which likely reflects observer effort (more people = more observations). One important caveat I wanted to be transparent about: this map reflects where people report bat sightings, not necessarily where bats live. iNaturalist data carries an inherent urban bias, more people means more observations. Coastal cities appear as hotspots partly because that’s where most California residents live. I acknowledged this directly in the map’s caption rather than letting the visualization imply a false objectivity.

When are they most active?

The seasonal pattern shows a clear peak in April, with strong activity continuing through fall. This spring surge aligns with bat breeding season and the emergence of insects after winter, exactly when insectivorous bats need maximum calories. Observations continue year-round, though, suggesting California’s mild climate supports active bat populations even in winter months.

Together, these patterns reveal that citizen science isn’t just documenting bat presence, it’s uncovering where conservation attention is needed and when communities are most likely to encounter these beneficial species.

Design Decisions: From Urban Focus to Statewide Story

My initial goal was to highlight bats in urban communities specifically to show city-dwellers that bats are right outside their windows. But the data had other plans. iNaturalist’s location data varies wildly in precision. Some observations include exact street addresses, while others are vague county-level descriptions like “San Diego County, CA.” When I filtered for observations containing “Los Angeles” or “San Diego,” I inadvertently captured national parks, wildlife refuges, and regional wilderness areas, hardly urban ecosystems. I tried text-matching exclusions, then considered spatial analysis tools to clip observations to census-defined urban boundaries, but each approach felt like forcing a narrative the data didn’t support. Given that only 67 truly urban observations remained after aggressive filtering, I chose to focus on the pattern strongly supported by the data rather than force a limited urban narrative that would have introduced significant bias.

I chose honesty over my initial vision. The 1,843 statewide observations tell a more compelling story anyway, bats thrive across California’s incredibly diverse landscapes, and that diversity itself is worth celebrating. Below I break down the visual process of each plot and how it went from Rstudio to Affinity Designer where all other design elements where incorporated, adjusted, or altered.

The following code shows the pre-data process done before making simple plots

#.........................load libraries.........................
library(tidyverse)
library(here)
library(lubridate)
library(sf)
library(spatstat)Ec
library(stars)
library(tigris) # for California state boundary
options(tigris_use_cache = TRUE) # to avoid re-downloading caches data

#..........................import data...........................

calbats <- read_csv(here("data", "observations.csv"))

# clean data 1018

bats_clean <- calbats %>% 
 filter(
    captive_cultivated == FALSE,
    # remove NAs
    !is.na(latitude),
    !is.na(longitude),
    !is.na(observed_on)
    ) %>% 
  mutate(
    date = ymd(observed_on),
    year = year(date),
    month = month(date, label = TRUE, abbr = FALSE),
# Relabel entries to reduce diversity count inflation (not all species have common names) 
    common_name = case_when(
    scientific_name == "Myotis"           ~ "Unidentified Mouse-eared Bat",
    scientific_name == "Vespertilionidae" ~ "Unidentified Evening Bat",
    scientific_name == "Vespertilionoidea"~ "Unidentified Vespertilionoid Bat",
    scientific_name == "Chiroptera"       ~ "Unidentified Bat",
    TRUE ~ common_name  # Keep all other common names as-is
  ))

Visualizing the Unidentified

One design challenge became an unexpected strength: 36% of observations aren’t identified to species level. Initially, I considered excluding these as “incomplete data.” But that would have erased 635 observations from people who spotted bats but lacked the expertise to distinguish a California Myotis from a Yuma Myotis. That’s exactly the kind of participation citizen science depends on. Instead, I relabeled them as “Unidentified Mouse-eared Bat” or “Unidentified Evening Bat” and visually distinguished them from confirmed species using grayish blue for unidentified, a medium purple for confirmed. This design choice transforms a limitation into transparency, showing both what we know for certain and the community engagement. Note: At this point my color theme wasn’t set yet, but I was experimenting with the BuPu color palette from the RColorBrewer package were I got these color hexcodes.

Initial Barplot before redesigning in Affinity Designer

# Top Species Observed plot
# Prepare top species data
top_species <- bats_clean %>%
  count(common_name, sort = TRUE) %>%
  slice_max(n, n = 10) %>%
  mutate(
    identified = !str_detect(common_name, "Unidentified"), # Create TRUE/FALSE column
    common_name = fct_reorder(common_name, n) # top observation at top
  ) %>% 
# Create bar plot
ggplot(aes(x = n, y = common_name, fill = identified)) +
  geom_col(alpha = 0.9) +
  scale_fill_manual(
    values = c("FALSE" = "#E0ECF4", "TRUE" = "#8856A7"),  # hexcodes from BuPu color palette 
    labels = c("Awaiting ID", "Identified to species"),
    name = NULL
  ) +
  labs(
    title = "Most documented Bat Species in California",
    subtitle = "Top 10 most observed species/groups",
    x = "Number of observations",
    y = NULL,
    caption = "64% of observations identified to species level"
  )  +
   theme_minimal(base_size = 12) +
  theme(
    plot.title = element_text(size = 14, face = "bold", hjust = 0),
    plot.subtitle = element_text(size = 11, color = "gray30"),
    plot.caption = element_text(size = 9, hjust = 0, color = "gray50"),
    legend.position = "top",
    panel.grid.major.y = element_blank(),      # Remove horizontal gridlines
    panel.grid.minor = element_blank(),
    axis.text = element_text(size = 10)
  )

print(top_species)

Map Iterations

The geographic visualization went through several iterations. Initially, I plotted bat sightings as individual points across California, a simple and direct approach; however. with thousands of observations, the visualization became overly dense, particularly in regions such as the Bay Area and Los Angeles. I then tried a Density contour plot that obscured the California state boundary. I almost settled with hexagonal binning that aggregates observations into equal-area cells, revealing hotspots but it was difficult to see high observations over low observations. After some feedback and discussing my vision with my colleagues I found the {spatstat} package, which calculates the relative concentration of observations across continuous space rather than binning points into arbitrary boundaries like counties or grids. The results is a smooth heatmap where warmer colors reflect higher observations density -giving the map a visual language that echos the infographics broader dawn-to-dusk color story. Getting there requires a few technical hurdles. The density calculation only works with projections that use meters as their unit, so I transformed all the data to California Albers (EPSG:3310) before converting observations to a spatial point pattern object (ppp). From there, density.pp() handled the math, and {stars} bridged the result back to a plottable {sf} object. I opt-out the legend because the raw density numbers (like 0.0000003) are meaningless to a general audience. Instead I added a text annotation in my infographic.

# Observation heatmap steps
# CRS (California Albers - uses meters, works with density .ppp)
ca_crs <- 3310

# Get California state boundary
ca_state <- states(cb = TRUE) %>% 
  filter(NAME == "California") %>% 
  st_transform(ca_crs)

# Convert bat observation to sf, drop Nas, clip to CA
bats_sf <- bats_clean %>% 
  filter(!is.na(latitude), !is.na(longitude)) %>% 
  st_as_sf(coords = c("longitude", "latitude"), crs = 4326) %>% 
  st_transform(ca_crs) %>% 
  st_intersection(ca_state) # remove any points outside CA

# Convert to ppp object 
bats_ppp <- as.ppp(bats_sf$geometry, W = as.owin(ca_state))

# Calculate density
bats_density_stars <- stars::st_as_stars(density(bats_ppp, dimyx = 150)) # change to 300 to save image 150 for faster output

# Convert back to sf
bats_density_sf <- st_as_sf(bats_density_stars) %>% 
  st_set_crs(ca_crs)

# Plot
camap <- ggplot() +
  geom_sf(data = bats_density_sf, aes(fill = v), color = NA) +
  geom_sf(data = ca_state, fill = NA, color = "white", linewidth = 0.35) +
  geom_sf(data = bats_sf, color = "white", alpha = 0.2, size = 0.5) +
  scale_fill_viridis_c(option = "magma", guide = "none") +
  theme_void()

print(camap)

Adapting the Radial Design

For the seasonal pattern, I chose a radial chart over a traditional bar chart. While bar charts are more familiar, the circular format reinforces that bat observations are cyclical. Spring leads to summer leads to fall lead back to spring. The April peak becomes visual prominent as the longest “spoke” on the wheel, and viewers can intuitively grasp the year-round presence of bats even during winter months. It also let me implement a creative design on my infographic by not just being a data plot but providing a thematic visual element.

Radial chart before

# Prepare monthly data
monthly <- bats_clean %>%
  count(month) %>%
  mutate(month = factor(month, levels = month.name))  # Keep month order

# Create radial chart
april <-  ggplot(monthly, aes(x = month, y = n)) +
  
  # BARS: Radial bars from center outward
  geom_col(fill = "#810f7c", width = 0.9) +
  
  # MAKE IT CIRCULAR: coord_polar converts x-axis to circle
  coord_polar(start = -pi/12) +  # Rotate so January starts at top
  
  # LABELS
  labs(
    title = "Seasonal observation patterns throughout the year",
    subtitle = "Peak activity in April"
  ) +
  
  # THEME: Clean radial appearance
  theme_minimal(base_size = 12) +
  theme(
    plot.title = element_text(size = 14, face = "bold", hjust = 0.5),
    plot.subtitle = element_text(size = 10, color = "gray30", hjust = 0.5),
    plot.caption = element_text(size = 9, hjust = 0.5, color = "gray50"),
    axis.title = element_blank(),           # Remove axis labels
    axis.text.y = element_blank(),          # Remove radial axis numbers
    axis.ticks = element_blank(),           # Remove tick marks
    panel.grid.major.x = element_line(color = "#b3cde3", linewidth = 0.3),# Month dividers
    panel.grid.major.y = element_blank(),   # Remove circular gridlines
    panel.grid.minor = element_blank()
  )
print(april)

Bringing it All Together: Typography, Color, and Accessibility

With three distinct visualizations built, the final design challenges was weaving them into a cohesive visual story. I brought all elements into Affinity Designer, where typography, color, and layout decisions were refined together rather than in isolation.

For typography, I paired Century Schoolbook for the titles with Garamond for body text and captions. Century Schoolbook’s sturdy, classic serifs give the title hierarchy weight and presence while Garamond’s elegant, lighter letter forms keep the supporting text readable without competing for attention. Both are serif fonts, which gives the infographic a consistent, editorial quality, fitting for a science communication piece.

Color decisions were made at the infographic level rather than the plot level. The dawn gradient background, transitioning from deep night sky to warm morning light, anchors the entire composition thematically and gives each visualization a natural home: the dark density map sits in both night and day, the radial sun glows at the dawn horizon.

Accessibility shaped several choices throughout. The magma palette on the density map and the orange-to-gold gradient on the radial chart both maintain strong contrast across their ranges. I avoided relying on color alone to communicate meaning, each plot uses position, labels, or annotations as a secondary channel.

Finally, framing bats as “hidden neighbors” rather than threats was itself an accessibility choice, making the subject approachable and empathy-building for audiences who might otherwise disengage.

Footnotes

1. FUJITA, M.S. and TUTTLE, M.D. (1991), Flying Foxes (Chiroptera: Pteropodidae): Threatened Animals of Key Ecological and Economic Importance. Conservation Biology, 5: 455-463. https://doi.org/10.1111/j.1523-1739.1991.tb00352.x

2. Kunz, T.H., Braun de Torrez, E., Bauer, D., Lobova, T. and Fleming, T.H. (2011), Ecosystem services provided by bats. Annals of the New York Academy of Sciences, 1223: 1-38. https://doi.org/10.1111/j.1749-6632.2011.06004.x

3. Russo, D. and Leonardo A. (2015), Sensitivity of bats to urbanization: a review. Mammalian Biology 80: 205-212. https://doi.org/10.1016/j.mambio.2014.10.003