Temperature Analysis Lab

Welcome to our Maunalua Fishpond Heritage Center Data Analysis Workshop 1!

In this R Markdown, we are going to:

• Clean and Visualize Temperature Data from Novemeber 2023 - February 2024 from 5 locations
  • Kalahaihai
    • Garage
  • Kanewai
    • Auwai Under Footbridge
    • Fishpond Edge Norfolk
    • Rock Stairs by Wall
    • Spring Ledge Makai End

Load Necessary Packages

We will begin by loading in the packages we will use in our analysis

• It may be helpful to think of packages as toolboxes that contain functions or tools that help us with our data analysis

Load Data into R

Our data is in the data folder in our R project

• Now we need to import it into R so we can view it

Here is the pseudo code

• what_we_want_to_name_the_dataframe <- read_csv(here(“folder_data_is_in/file_name.csv))
  • read_csv() is a function from the tidyverse package that - reads files ending in .csv
  • here() cuts out long file paths so we don’t have to set working directories, and it makes our code reproducible

temp_kalauhaihai_garage <- read_csv(here("data/data_temp_NovFeb2024_KalauhaihaiGarage_20970109.csv"), skip = 1)

temp_kanewai_auwai_under_footbridge <- read_csv(here("data/data_temp_NovFeb2024_KanewaiAuwaiUnderFootBridge_21445021.csv"), skip = 1)

temp_kanewai_fishpond_edge_norfolk <- read_csv(here("data/data_temp_NovFeb2024_KanewaiFishpondEdgeNorfolk_20970123.csv"), skip = 1)

temp_kanewai_rockstairs_by_wall <- read_csv(here("data/data_temp_NovFeb2024_KanewaiRockStairsByWall_20970115.csv"), skip = 1)

temp_kanewai_spring_ledge_makai_end <- read_csv(here("data/data_temp_NovFeb2024_KanewaiSpringLedgeMakaiEnd_20970094.csv"), skip = 1)

Aweosme! You can now view the data by clicking on it in your environment (top right pane in R Studio) or by writing code like this to view the first 10 rows

head(temp_kalauhaihai_garage)

## # A tibble: 6 × 8
##     `#` `Date Time, GMT-10:00` Temp, °C (LGR S/N: 20970…¹ Coupler Detached (LG…²
##   <dbl> <chr>                                       <dbl> <chr>                 
## 1     1 11/02/23 01:00:00                            22.3 Logged                
## 2     2 11/02/23 01:10:00                            22.2 <NA>                  
## 3     3 11/02/23 01:20:00                            22.2 <NA>                  
## 4     4 11/02/23 01:30:00                            22.2 <NA>                  
## 5     5 11/02/23 01:40:00                            22.3 <NA>                  
## 6     6 11/02/23 01:50:00                            22.2 <NA>                  
## # ℹ abbreviated names: ¹​`Temp, °C (LGR S/N: 20970109, SEN S/N: 20970109)`,
## #   ²​`Coupler Detached (LGR S/N: 20970109)`
## # ℹ 4 more variables: `Coupler Attached (LGR S/N: 20970109)` <lgl>,
## #   `Host Connected (LGR S/N: 20970109)` <lgl>,
## #   `Stopped (LGR S/N: 20970109)` <chr>,
## #   `End Of File (LGR S/N: 20970109)` <chr>

Cleaning our Data

colnames(temp_kalauhaihai_garage)

## [1] "#"                                              
## [2] "Date Time, GMT-10:00"                           
## [3] "Temp, °C (LGR S/N: 20970109, SEN S/N: 20970109)"
## [4] "Coupler Detached (LGR S/N: 20970109)"           
## [5] "Coupler Attached (LGR S/N: 20970109)"           
## [6] "Host Connected (LGR S/N: 20970109)"             
## [7] "Stopped (LGR S/N: 20970109)"                    
## [8] "End Of File (LGR S/N: 20970109)"

You’ll notice the column names in our dataframes have capital letters and are seperated by spaces. This can cause errors in R, so we change all of our column names to lowercase letters separating words with underscores (ex. Date Time, GMT-10:00 to date_time)

• We can do this with the clean_names() function from the janitor package

The pseudo code here is very common:

• new_dataframe_name <- original_dataframe_name %>% action_to_data()
  • in this case action_to_data will be clean_names()
  • also, if you want to overwrite the dataframe (only of your not making big changes) you can just keep the new_dataframe_name the same as the original_dataframe_name

temp_kalauhaihai_garage <- temp_kalauhaihai_garage %>% 
  clean_names()

temp_kanewai_auwai_under_footbridge <- temp_kanewai_auwai_under_footbridge %>% 
  clean_names()

temp_kanewai_fishpond_edge_norfolk <- temp_kanewai_fishpond_edge_norfolk %>% 
  clean_names()

temp_kanewai_rockstairs_by_wall <- temp_kanewai_rockstairs_by_wall %>% 
  clean_names()

temp_kanewai_spring_ledge_makai_end <- temp_kanewai_spring_ledge_makai_end %>% 
  clean_names()

Now let’s check out the difference

colnames(temp_kalauhaihai_garage)

## [1] "number"                                  
## [2] "date_time_gmt_10_00"                     
## [3] "temp_c_lgr_s_n_20970109_sen_s_n_20970109"
## [4] "coupler_detached_lgr_s_n_20970109"       
## [5] "coupler_attached_lgr_s_n_20970109"       
## [6] "host_connected_lgr_s_n_20970109"         
## [7] "stopped_lgr_s_n_20970109"                
## [8] "end_of_file_lgr_s_n_20970109"

Great! These are now much easier to work with, but still very long

• Let’s shorten
  • temp_c_lgr_s_n_20970109_sen_s_n_20970109 to temp_celcius
  • date_time_gmt_10_00 to date_time

colnames(temp_kalauhaihai_garage)[colnames(temp_kalauhaihai_garage) == "date_time_gmt_10_00"] = "date_time"

colnames(temp_kanewai_auwai_under_footbridge)[colnames(temp_kanewai_auwai_under_footbridge) == "date_time_gmt_10_00"] ="date_time"

colnames(temp_kanewai_fishpond_edge_norfolk)[colnames(temp_kanewai_fishpond_edge_norfolk) == "date_time_gmt_10_00"] ="date_time"

colnames(temp_kanewai_rockstairs_by_wall)[colnames(temp_kanewai_rockstairs_by_wall) == "date_time_gmt_10_00"] ="date_time"

colnames(temp_kanewai_spring_ledge_makai_end)[colnames(temp_kanewai_spring_ledge_makai_end) == "date_time_gmt_10_00"] ="date_time"

Be careful with temperature because every dataframe has a slightly different column name for temeprature

colnames(temp_kalauhaihai_garage)[colnames(temp_kalauhaihai_garage) == "temp_c_lgr_s_n_20970109_sen_s_n_20970109"] ="temp_celcius"

colnames(temp_kanewai_auwai_under_footbridge)[colnames(temp_kanewai_auwai_under_footbridge) == "temp_c_lgr_s_n_21445021_sen_s_n_21445021"] ="temp_celcius"

colnames(temp_kanewai_fishpond_edge_norfolk)[colnames(temp_kanewai_fishpond_edge_norfolk) == "temp_c_lgr_s_n_20970123_sen_s_n_20970123"] ="temp_celcius"

colnames(temp_kanewai_rockstairs_by_wall)[colnames(temp_kanewai_rockstairs_by_wall) == "temp_c_lgr_s_n_20970115_sen_s_n_20970115"] ="temp_celcius"

colnames(temp_kanewai_spring_ledge_makai_end)[colnames(temp_kanewai_spring_ledge_makai_end) == "temp_c_lgr_s_n_20970094_sen_s_n_20970094"] ="temp_celcius"

Let’s check the results

colnames(temp_kalauhaihai_garage)

## [1] "number"                            "date_time"                        
## [3] "temp_celcius"                      "coupler_detached_lgr_s_n_20970109"
## [5] "coupler_attached_lgr_s_n_20970109" "host_connected_lgr_s_n_20970109"  
## [7] "stopped_lgr_s_n_20970109"          "end_of_file_lgr_s_n_20970109"

Perfect! Now let’s select the columns that are relevant to us using the select function

• date_time
• temp_celcius

temp_kalauhaihai_garage <- temp_kalauhaihai_garage %>% 
  select(date_time, temp_celcius)

temp_kanewai_auwai_under_footbridge <- temp_kanewai_auwai_under_footbridge %>% 
  select(date_time, temp_celcius)

temp_kanewai_fishpond_edge_norfolk <- temp_kanewai_fishpond_edge_norfolk %>% 
  select(date_time, temp_celcius)

temp_kanewai_rockstairs_by_wall <- temp_kanewai_rockstairs_by_wall %>% 
  select(date_time, temp_celcius)

temp_kanewai_spring_ledge_makai_end <- temp_kanewai_spring_ledge_makai_end %>% 
  select(date_time, temp_celcius)

Let’s check the results

head(temp_kalauhaihai_garage)

## # A tibble: 6 × 2
##   date_time         temp_celcius
##   <chr>                    <dbl>
## 1 11/02/23 01:00:00         22.3
## 2 11/02/23 01:10:00         22.2
## 3 11/02/23 01:20:00         22.2
## 4 11/02/23 01:30:00         22.2
## 5 11/02/23 01:40:00         22.3
## 6 11/02/23 01:50:00         22.2

Every object (something we assign a name to) has a class in R

The class of temp_kalauhaihai_garage is a dataframe

class(temp_kalauhaihai_garage)

## [1] "tbl_df"     "tbl"        "data.frame"

The class of the date_time column in temp_kalauhaihai_garage is a character

class(temp_kalauhaihai_garage$date_time)

## [1] "character"

We need R to recognize that this is a date, so we need to manually change the class using the mdy_hms() function from the lubridate package

• mdy_hms stand for month day year :: hour minute second
• this is how the date is currently formatted

temp_kalauhaihai_garage$date_time <- mdy_hms(temp_kalauhaihai_garage$date_time)

temp_kanewai_auwai_under_footbridge$date_time <- mdy_hms(temp_kanewai_auwai_under_footbridge$date_time)

temp_kanewai_fishpond_edge_norfolk$date_time <- mdy_hms(temp_kanewai_fishpond_edge_norfolk$date_time)

temp_kanewai_rockstairs_by_wall$date_time <- mdy_hms(temp_kanewai_rockstairs_by_wall$date_time)

temp_kanewai_spring_ledge_makai_end$date_time <- mdy_hms(temp_kanewai_spring_ledge_makai_end$date_time)

Let’s check the results (POSIXct is equivalent to date/time)

class(temp_kalauhaihai_garage$date_time)

## [1] "POSIXct" "POSIXt"

The last step before visualizing our temperature data is trimming the data to only include the time the sensor was in the water at each given location

Let’s make a plot to see what the data looks like untrimmed using the ggplot() function

the pseudo code for this is

• ggplot(data = name_of_dataframe_you_want_to_plot, aes(x = column_you_want_on_the_x_axis, y = column_you_want_on_the_y_axis)) + geom_whatever_kind_of_chart_you_want_to_make()

ggplot(data = temp_kalauhaihai_garage, aes(x = date_time,
                                          y = temp_celcius)) +
  geom_line()

Let’s store and plot the start and end times Dr. Lupita has recorded as variables for the kalauhaihai garage sensor

• start time: 11/2/2023 9:30:00
• end time: 1/31/2024 9:33:00

kalauhaihai_garage_start_time <- as.POSIXct("11-02-2023 9:30:00", format = "%m-%d-%Y %H:%M:%S")

kalauhaihai_garage_end_time <- as.POSIXct("01-31-2024 9:33:00", format = "%m-%d-%Y %H:%M:%S")

ggplot(data = temp_kalauhaihai_garage, aes(x = date_time,
                                          y = temp_celcius)) +
  geom_line() +
  geom_vline(xintercept = as.numeric(kalauhaihai_garage_start_time), linetype = "dashed", color = "darkgreen") +
  geom_vline(xintercept = as.numeric(kalauhaihai_garage_end_time), linetype = "dashed", color = "darkred")

Let’s make a new dataframe called temp_kalauhaihai_garage_trimmed with the trimmed data

temp_kalauhaihai_garage_trimmed <- temp_kalauhaihai_garage %>%
  filter(date_time >= kalauhaihai_garage_start_time & date_time <= kalauhaihai_garage_end_time)

Let’s plot the trimmed data (update the dataframe name)

ggplot(data = temp_kalauhaihai_garage_trimmed, aes(x = date_time,
                                          y = temp_celcius)) +
  geom_line()

Awesome! Now let’s cutomize the plot and make it interactive

• Assign the plot a name with <-
• Change the color
• Add title and change axis labels with labs()
• Change the theme with theme_minimal()
• Make interactive with ggplotly()

temp_kalauhaihai_garage_plot <- ggplot(data = temp_kalauhaihai_garage_trimmed, aes(x = date_time,
                                          y = temp_celcius)) +
  geom_line(color = "steelblue") +
  theme_minimal() +
  labs(title = "Temperature Kalauhaihai Garage",
       subtitle = "November 2023 - February 2024",
       x = "Date",
       y = "Temperature °C")

ggplotly(temp_kalauhaihai_garage_plot)

Question

• What happened on December 14th?

Challenge

Can you

1. Trim the data for the other 4 locations
2. Visualize them