#Install packages needed
pacman::p_load(
# Package Install and Management
pacman, # package install/load
janitor, #clean up data names
# Project and File Management
readr, # import data
httr, #github passkey
usethis, #r environment
RPostgres, #SQL
DBI, #SQL
jsonlite, #working with json
rvest, #html files
rstuioapi, #api
knitr, #displaying data in quarto
# General Data Management
tidyverse, # work with dates
scales, # ggplot
gmodels, # freq tables SAS
skimr,
hms, # recommended for dataset 4
patchwork #data visualization, side by side
)Class 607, Project 2 Data Transformations
Instructions here: https://brightspace.cuny.edu/d2l/le/content/1404182/viewContent/41373795/View
Introduction
Data sources: All datasets originated from students in Discussion #5a.
Dataset 1 from student Patricio Romero: “For this discussion, I selected the Electric Consumption and Cost dataset from NYC Open Data. This dataset contains electrical energy information from New York City Housing Authority (NYCHA) buildings.
Analysis request: compare electricity consumption between NYC boroughs and see how electricity use changes over time.
Data Source: NYC Open Data – Electric Consumption And Cost (2010–Sep 2025)
Dataset Link: https://data.cityofnewyork.us/Housing-Development/Electric-Consumption-And-Cost-2010-Sep-2025-/jr24-e7cr
Dataset 2 from student Renee Watson: “fictional dataset containing monthly patient visits for community health clinics during 2025
Analysis request: I would examine how patient visit volume changes throughout the year and compare these patterns across the five clinics. I could determine which clinics experienced the largest increases or decreases in patient visits and examine whether there are noticeable monthly patterns in patient volume. A visualization could also be used to compare trends across the clinics.”
Dataset 3 from student Kingsley Kpoku: “Fictional NYC housing prices dataset for the five New York Boroughs .The dataset is on 2024 and 2025.
Analysis request: The whole idea is to analyze and study the behavioral patterns of the averages prices across the months and Years and serve as a real estate analyst to advice potential customers on the most expensive areas in NewYork, what people can afford based on salary and when is the best time to move to a borough when the prices drop (based on historicals)”
*Dataset 4 from student Noelle Messaoudi: “As a single mom who works part-time and attends graduate school part-time, I often have to fit my workload and studying around my toddler’s sleep. This gave me the idea to create a fictional dataset based on my experience of trying to manage everything in one day.
The dataset would record a child’s sleep over two weeks. It would be untidy because each day’s nap and bedtime would have separate columns, instead of having one row for each day.
Analysis request: For Project 2, I could explore whether longer naps are associated with later bedtimes and how much the sleep schedule varies from day to day.I am curious about this because “I will study when he sleeps” sounds like a plan, until he decides otherwise.”
*adding one more since Dataset 1 wasn’t a wide dataset. Since this data is an example, I used Claude to expand on it.
Tackling the problem: In all three datasets I’ll need to extract the wide, messy data in different ways from online (dataset #1), the discussion body (dastaset #2), or downloaded (dataset #3). Then, I’ll import them all in R and begin reshaping and I’ll review the different analysis requests to determine what more is needed for preparation. Finally, I’ll tackle the analysis questions with a summary table, visualization, and short narrative.
Anticipated data challenges: Dataset1 seems like it will be the most challenging because I don’t know anything about how to analyize energy consumption, and the timespan is 15 years, so it requires deduplicating and sifting through a lot of data. I also anticipate all the datacleaning across the three datasets to be challenging.
Citations:
- Anthropic. (2025). Claude Opus 4.5 [Large language model]. (https://claude.ai/) Accessed November 8, 2026. Link to chat.
Analysis
Load Packages
Load Data
First, I am deduplicating and dropping NAs to dataset1 so it can be uploaded to github.
#locations
#input_location <- "~/Downloads/"
#location_desktop <- "~/Desktop/"
#dataset1 <- readr::read_csv(paste0(location_desktop, "dataset1.csv"), show_col_types = FALSE)
#look at date variable
#skim(dataset1$`Revenue Month`)
#prepare for creating a subset with unique bill ids and no na
#dataset1_sub <- dataset1 %>%
# clean_names() %>%
# distinct(umis_bill_id, .keep_all = TRUE)%>%
# drop_na()
#check date range
#range(dataset1_sub$revenue_month)
#create subset file
#write_csv(dataset1_sub, file.path(location_desktop, paste0("dataset1_sub.csv")))Next, load data saved as .csv in github and import into R
location <- "~/Desktop/CUNY/607-assignments/projects/project 2/"
#Dataset 1:
dataset1 <- read_csv(paste0(location, "dataset1_sub.csv"), show_col_types = FALSE)
#Dataset 2:
dataset2 <- read_csv(paste0(location, "dataset2.csv"), show_col_types = FALSE)
#Dataset 3:
dataset3 <- read_csv(paste0(location, "dataset3.csv"), show_col_types = FALSE)
#Dataset 4:
dataset4 <- read_csv(paste0(location, "dataset4.csv"), show_col_types = FALSE)#
#
Dataset 1
Data cleaning
#create additional variables that we need for the analysis.
dataset1_sub <- dataset1 %>%
mutate(
revenue_month = ym(revenue_month),
building_id = paste(coalesce(as.character(tds_number), as.character(edp)),
location, sep = "-"),
# use per day, since bills cover different numbers of days
consumption_per_day = consumption_kwh / number_days)
#take care of invalid, NAs, duplicates (see checks)
#Keep the highest bill ID per building, meter and revenue month
dataset1_2sub <- dataset1_sub%>%
group_by(building_id, meter_number, revenue_month) %>%
slice_max(umis_bill_id, n = 1, with_ties = FALSE) %>%
#drop nas and any 0 values
drop_na(borough, revenue_month, consumption_kwh, number_days) %>%
filter(consumption_kwh > 0, number_days > 0) %>%
ungroup()
#now create dataset and add variables
dataset1_3sub <- dataset1_2sub |>
#add number of buildings per borough and total consumption
group_by(borough, revenue_month) %>%
summarize(
n_buildings = n_distinct(building_id),
mo_total_kwh = sum(consumption_kwh),
mo_kwh_per_day = sum(consumption_per_day),
.groups = "drop"
) %>%
mutate(
kwh_per_bldg_per_day = mo_kwh_per_day / n_buildings,
# date and season
rev_year = year(revenue_month),
rev_month = month(revenue_month),
rev_season = case_when(rev_month %in% c(12, 1, 2) ~ "winter",
rev_month %in% 3:5 ~ "spring",
rev_month %in% 6:8 ~ "summer",
rev_month %in% 9:11 ~ "fall"),
# year-over-year change on that measure
kwh_last_year = lag(kwh_per_bldg_per_day, 12),
yoy_pct_change = (kwh_per_bldg_per_day - kwh_last_year) / kwh_last_year * 100)%>%
ungroup()
#CHECKS FOR DUPLICATES
dataset1_2sub %>% count(umis_bill_id) %>% count(n, name = "n_bill_ids")
#checking building rows per month
dataset1_2sub %>%
count(building_id, revenue_month) %>%
count(n, name = "n_building_months")
#should be none
dup_check <- dataset1_2sub %>%
group_by(building_id, meter_number, revenue_month) %>%
filter(n() > 1) %>%
arrange(building_id, meter_number, revenue_month)Run analysis
Compare electricity consumption between NYC boroughs and see how electricity use changes over time.
Table:
annual_long <- dataset1_3sub %>%
# 2025 only has data up to Sept so exclude. Also exclude non development city as a borough
filter(rev_year < 2025, borough != "NON DEVELOPMENT FACILITY") %>%
group_by(borough, rev_year) %>%
summarize(avg_kwh = mean(kwh_per_bldg_per_day, na.rm = TRUE), .groups = "drop")
#create table
annual_wide <- annual_long %>%
select(rev_year, borough, avg_kwh)%>%
pivot_wider(names_from = c(borough), values_from = avg_kwh)
# display in quarto
kable(annual_wide,
col.names = c("Year", "Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island"),
caption = "Annual Average Daily Electricity Use (kWh) per Building, 2010 - 2024") | Year | Bronx | Brooklyn | Manhattan | Queens | Staten Island |
|---|---|---|---|---|---|
| 2010 | 1455.492 | 186.9330 | 462.2863 | 2822.082 | 636.0199 |
| 2012 | 1360.134 | 191.0044 | 506.3698 | 2603.030 | 714.3561 |
| 2013 | 1294.160 | 192.3820 | 531.2423 | 2469.217 | 605.7245 |
| 2014 | 1220.432 | 198.1961 | 517.3521 | 2461.666 | 573.8604 |
| 2015 | 1221.958 | 194.1154 | 1117.9037 | 2132.105 | 1384.7514 |
| 2016 | 1189.571 | 183.7798 | 1057.4120 | 2378.037 | 1280.1470 |
| 2017 | 1142.234 | 180.3669 | 1011.2828 | 2397.003 | 1298.2257 |
| 2019 | 1111.510 | 176.3655 | 983.2447 | 2239.892 | 1129.9895 |
| 2020 | 1185.011 | 162.6253 | 944.8685 | 1912.892 | 1295.4225 |
| 2021 | 1188.835 | 160.9826 | 933.4759 | 1857.763 | 1183.4413 |
| 2022 | 1179.243 | 176.1399 | 895.3775 | 1877.424 | 1157.8735 |
| 2023 | 1619.539 | 1159.6294 | 365.1579 | 1065.870 | 1244.7210 |
| 2024 | 1061.356 | 168.8994 | 839.2531 | 1465.844 | 1208.2962 |
Graph:
ggplot(annual_long, aes(rev_year, avg_kwh, color = borough)) +
geom_line(linewidth = 0.9) + geom_point() +
scale_y_continuous(labels = comma) +
labs(title = "Annual Average Daily Electricity Use per Building 2010 - 2024",
x = "Year", y = "kWh per building per day", color = "Borough") +
theme_minimal()
Summary
Overall we see that yearly average energy consumption (KWh per building/day) is different by borough. Notably Queens has steadily decreased from approximately 2750 in 2010 to 1500 in 2024. in Brooklyn, energy consumption has stayed low until 2023 and then spiked up above 1000 in 2023.
#
#
Dataset 2
Data cleaning
#clean headers
dataset2_sub <- dataset2 %>% clean_names()
#transform the dataset from wide to long to see all dates in one column
dataset2_long <- dataset2_sub %>%
pivot_longer(
cols = !clinic,
names_to = "month",
values_to = "count")%>%
#edit month to date variables and add percents
mutate(
#month_num = match(str_to_title(str_sub(month, 1, 3)), month.abb),
month = factor(str_to_title(str_sub(month, 1, 3)), levels = month.abb),
totals = sum(count), .by= clinic,
pct_total = round((count / totals) *100, 1))Run analysis
How do patient visit volume changes throughout the year? Compare these patterns across the five clinics.
Table:
#display work with this
dataset2_long|>
select(clinic, month, pct_total)%>%
kable(col.names = c("clinic", "month", "pct_total"))
#create table
dataset2_table <- dataset2_long %>%
select(clinic, month, pct_total)%>%
pivot_wider(names_from = c(clinic), values_from = pct_total)# display in quarto
kable(dataset2_table,
caption = "Monthly Percent of Patient Visit Volume, by Clinic") | month | Harbor Community Health Center | Riverside Family Health Clinic | Central Community Health Center | Eastside Family Clinic | Westview Community Health Center | Northside Health Clinic | Greenwood Family Health Center | Lakeside Community Clinic | Parkview Health Center | Hillcrest Community Clinic |
|---|---|---|---|---|---|---|---|---|---|---|
| Jan | 7.1 | 6.6 | 7.2 | 6.7 | 7.1 | 7.0 | 6.7 | 7.2 | 6.9 | 6.6 |
| Feb | 6.7 | 7.1 | 6.9 | 7.0 | 6.9 | 6.8 | 6.6 | 7.0 | 7.1 | 6.8 |
| Mar | 7.5 | 6.9 | 7.6 | 7.3 | 7.3 | 7.4 | 7.2 | 7.4 | 7.4 | 7.3 |
| Apr | 8.1 | 7.7 | 7.9 | 7.2 | 7.7 | 7.8 | 7.7 | 7.8 | 7.7 | 7.6 |
| May | 8.4 | 8.2 | 8.3 | 7.8 | 8.1 | 8.1 | 8.1 | 8.0 | 8.1 | 7.9 |
| Jun | 8.7 | 8.4 | 8.1 | 8.4 | 8.4 | 8.4 | 8.2 | 8.4 | 8.4 | 8.2 |
| Jul | 9.1 | 8.7 | 8.6 | 8.7 | 8.3 | 8.7 | 8.6 | 8.6 | 8.2 | 8.6 |
| Aug | 9.0 | 9.1 | 8.5 | 9.1 | 8.7 | 8.5 | 8.9 | 8.5 | 8.7 | 8.8 |
| Sep | 8.3 | 8.9 | 8.8 | 8.8 | 9.0 | 9.0 | 9.1 | 8.9 | 9.0 | 9.1 |
| Oct | 8.6 | 9.4 | 9.1 | 9.4 | 9.2 | 9.2 | 9.4 | 9.2 | 9.2 | 9.4 |
| Nov | 9.0 | 9.3 | 9.3 | 9.7 | 9.4 | 9.4 | 9.6 | 9.4 | 9.6 | 9.6 |
| Dec | 9.4 | 9.7 | 9.7 | 9.9 | 9.8 | 9.8 | 9.9 | 9.7 | 9.8 | 10.0 |
Graph:
ggplot(dataset2_long, aes(month, pct_total, color = clinic, group = clinic)) +
geom_line(linewidth = 0.9) + geom_point() +
labs(title = "Monthly Patient Visit Volume by Clinic in 2025",
x = "Month", y = "Percent of clinic's yearly total", color = "Clinic") +
theme_minimal()
Summary
At all clinics, patient visits saw an upward trend from 7% to 10% by the end of the year, with a slight dip in February. Harbor Community Health Center saw the highest volume of patient visits from March to July then switched to the lowest visit percentage from September to December.
#
#
Dataset 3
Data cleaning
#clean headers
dataset3_sub <- dataset3 %>% clean_names()
#weird formatting split the dataset into two parts to fix
first_part <- dataset3_sub %>% slice(1:24)
second_part <- dataset3_sub %>% slice(25:n())
#fix second part
second_part_wide <- second_part%>%
select(year, month, bronx, brooklyn) %>%
pivot_wider(
names_from = bronx,
values_from = brooklyn)
#reclean headers
second_part_wide <- second_part_wide%>% clean_names()
#change bronx to numeric in both parts
first_part$bronx <- as.numeric(first_part$bronx)
second_part_wide$bronx <- as.numeric(second_part_wide$bronx)
#put back together with first part
dataset3_clean <- bind_rows(first_part, second_part_wide)
#edit and add variables
dataset3_clean <- dataset3_clean %>%
mutate(
month = factor(str_to_title(str_sub(month, 1, 3)), levels = month.abb),
date = make_date(year, month, 1),
year = format(date, "%Y"),
date_label = format(date, "%Y-%m"))%>%
select(-c(month, date))
#pivot longer
dataset3_long <- dataset3_clean %>%
pivot_longer(
cols = -c(date_label,year),
names_to = "borough",
values_to = "price")%>%
# keep one row per borough-month
arrange(borough, date_label, is.na(price)) %>%
distinct(borough, date_label, .keep_all = TRUE) %>%
#year total by borough
mutate(
yr_totals = sum(price, na.rm = TRUE),
yr_pct_total = round(price / yr_totals * 100, 1),
.by = c(borough, year)
) %>%
# month total
mutate(
mo_totals = sum(price, na.rm = TRUE),
mo_pct_total = round(price / mo_totals * 100, 1),
.by = date_label
)%>%
#make date_label back to date
mutate(
date_label = ym(date_label)
)Run analysis
What are the average prices across the months and years in each NYC borough? Is there a best time to move to a borough when the prices drop?
Table:
#create table
dataset3_sum <- dataset3_long %>%
summarise(
avg_price = round(mean(price), 1),
#avg_mo_pct = round(mean(mo_pct_total),1),
n = sum(!is.na(price)),
.by = c(year, borough))%>%
arrange(year, borough)%>%
select(-n)
# display in quarto
kable(dataset3_sum,
col.names = c("Year", "Borough", "Avg Price"),
caption = "Yearly Average Real Estate Prices, by Borough") | Year | Borough | Avg Price |
|---|---|---|
| 2024 | bronx | 580.6 |
| 2024 | brooklyn | 975.0 |
| 2024 | manhattan | 1130.4 |
| 2024 | queens | 669.7 |
| 2024 | staten_island | 695.6 |
| 2025 | bronx | 608.7 |
| 2025 | brooklyn | 1025.2 |
| 2025 | manhattan | 1174.8 |
| 2025 | queens | 704.0 |
| 2025 | staten_island | 733.4 |
Graph:
ggplot(dataset3_long, aes(date_label, mo_pct_total, color = borough, group = borough)) +
geom_line(linewidth = 0.9) + geom_point() +
scale_x_date(date_breaks = "3 months", date_label = "%Y-%m") +
labs(title = "Monthly Percent of Average Real Estate Prices, by Borough",
x = "Year-Month", y = "Percent of Average Real Estate Prices", color = "Borough") +
theme_minimal()
Summary
The average prices across each NYC borough from 2024-25 stayed relatively the same across time, but there were differences by borough. The most expensive average price was manhattan, followed by brooklyn. The lowest average place was in the bronx.
#
#
Dataset 4
Data cleaning
#clean headers
dataset4_sub <- dataset4 %>% clean_names()
#transform the dataset from wide to long to see all dates in one column
dataset4_long <- dataset4_sub %>%
pivot_longer(
cols = matches("^week_\\d+$"),
names_to = "week",
names_transform = readr::parse_number,
values_to = "time",
values_transform = as.character)%>%
#day_measure combines both measures, separate
separate(day_measure, into = c("day", "measure"), sep = " ", extra = "merge")%>%
pivot_wider(names_from = measure,
values_from = time)%>%
clean_names()
#add and edit time variables, determine wake time and sleep time. Package recommendation was hms for time mutations
dataset4_long <- dataset4_long %>%
mutate(
# "No nap" becomes NA
nap_start = na_if(nap_start, "No nap"),
nap_min = as.numeric(nap_min),
nap_hrs = round(nap_min / 60, 2),
day = factor(day, levels = c("Monday", "Tuesday", "Wednesday", "Thursday",
"Friday", "Saturday", "Sunday")),
# actual time-of-day values (hms class, prints as 12:45:00 / 19:30:00)
nap_start_time = as_hms(parse_date_time(nap_start, "I:M p")),
bedtime_time = as_hms(parse_date_time(bedtime, "I:M p")),
# 24-hour text labels, e.g. "12:45" and "19:30"
nap_start_24 = format(parse_date_time(nap_start, "I:M p"), "%H:%M"),
bedtime_24 = format(parse_date_time(bedtime, "I:M p"), "%H:%M"),
# nap end time = start + length (nap_min)
nap_end_time = as_hms(as.numeric(nap_start_time) + nap_min * 60),
awake_after_nap_min = as.numeric(bedtime_time - nap_end_time) / 60,
awake_after_nap_hr = round(awake_after_nap_min / 60, 2)
) Run analysis
Are longer naps are associated with later bedtimes/staying up longer?
Table:
dataset4_year_average <- dataset4_long %>%
group_by() %>%
summarize(
avg_nap_hrs = round(mean(nap_hrs, na.rm = TRUE), 2),
avg_awake_hrs = round(mean(awake_after_nap_hr, na.rm = TRUE), 2),
.groups = "drop"
)%>%
select(avg_nap_hrs, avg_awake_hrs )
dataset4_year_average#center around average year nap hrs and awake hours
#calculate averages
dataset4_average <- dataset4_long %>%
group_by(week) %>%
summarize(
avg_nap_hrs = round(mean(nap_hrs, na.rm = TRUE), 2),
avg_awake_hrs = round(mean(awake_after_nap_hr, na.rm = TRUE), 2),
.groups = "drop"
)%>%
mutate(
nap_centered = round((avg_nap_hrs - 1.73) / sd(avg_nap_hrs), 2),
awake_centered = round((avg_awake_hrs - 4.82) / sd(avg_awake_hrs), 2)
)%>%
select(week, avg_nap_hrs, avg_awake_hrs, nap_centered, awake_centered )
kable(dataset4_average,
col.names = c("Week (in calendar year)", "Avg nap (hrs)", "Avg awake after nap (hrs)", "Avg nap centered", "Avg awake after nap centered"),
caption = "Average nap length and awake time before bed, by week")| Week (in calendar year) | Avg nap (hrs) | Avg awake after nap (hrs) | Avg nap centered | Avg awake after nap centered |
|---|---|---|---|---|
| 1 | 2.18 | 4.61 | 1.82 | -1.18 |
| 2 | 1.93 | 4.46 | 0.81 | -2.03 |
| 3 | 2.11 | 4.79 | 1.54 | -0.17 |
| 4 | 2.04 | 4.64 | 1.25 | -1.01 |
| 5 | 1.93 | 4.64 | 0.81 | -1.01 |
| 6 | 2.11 | 4.64 | 1.54 | -1.01 |
| 7 | 1.93 | 4.86 | 0.81 | 0.23 |
| 8 | 2.04 | 4.79 | 1.25 | -0.17 |
| 9 | 2.07 | 4.54 | 1.37 | -1.58 |
| 10 | 1.61 | 4.88 | -0.48 | 0.34 |
| 11 | 2.00 | 4.79 | 1.09 | -0.17 |
| 12 | 2.18 | 4.50 | 1.82 | -1.80 |
| 13 | 2.00 | 4.57 | 1.09 | -1.41 |
| 14 | 1.96 | 4.71 | 0.93 | -0.62 |
| 15 | 1.82 | 4.89 | 0.36 | 0.39 |
| 16 | 1.75 | 4.83 | 0.08 | 0.06 |
| 17 | 1.64 | 4.75 | -0.36 | -0.39 |
| 18 | 1.57 | 5.00 | -0.65 | 1.01 |
| 19 | 1.54 | 5.08 | -0.77 | 1.46 |
| 20 | 1.75 | 4.58 | 0.08 | -1.35 |
| 21 | 1.86 | 4.75 | 0.53 | -0.39 |
| 22 | 1.54 | 5.04 | -0.77 | 1.24 |
| 23 | 1.71 | 4.86 | -0.08 | 0.23 |
| 24 | 1.71 | 5.00 | -0.08 | 1.01 |
| 25 | 1.93 | 4.82 | 0.81 | 0.00 |
| 26 | 1.79 | 4.75 | 0.24 | -0.39 |
| 27 | 1.86 | 4.86 | 0.53 | 0.23 |
| 28 | 1.57 | 5.00 | -0.65 | 1.01 |
| 29 | 1.39 | 4.88 | -1.37 | 0.34 |
| 30 | 1.61 | 4.75 | -0.48 | -0.39 |
| 31 | 1.86 | 4.71 | 0.53 | -0.62 |
| 32 | 1.36 | 4.65 | -1.50 | -0.96 |
| 33 | 1.64 | 4.62 | -0.36 | -1.13 |
| 34 | 1.86 | 4.50 | 0.53 | -1.80 |
| 35 | 1.68 | 5.04 | -0.20 | 1.24 |
| 36 | 1.89 | 4.89 | 0.65 | 0.39 |
| 37 | 1.46 | 4.83 | -1.09 | 0.06 |
| 38 | 1.36 | 4.65 | -1.50 | -0.96 |
| 39 | 1.43 | 5.04 | -1.21 | 1.24 |
| 40 | 1.64 | 4.75 | -0.36 | -0.39 |
| 41 | 1.43 | 4.96 | -1.21 | 0.79 |
| 42 | 1.61 | 5.21 | -0.48 | 2.20 |
| 43 | 1.14 | 4.80 | -2.38 | -0.11 |
| 44 | 1.79 | 4.79 | 0.24 | -0.17 |
| 45 | 1.82 | 4.89 | 0.36 | 0.39 |
| 46 | 2.00 | 4.82 | 1.09 | 0.00 |
| 47 | 1.54 | 4.92 | -0.77 | 0.56 |
| 48 | 1.68 | 5.11 | -0.20 | 1.63 |
| 49 | 1.36 | 4.92 | -1.50 | 0.56 |
| 50 | 1.64 | 5.04 | -0.36 | 1.24 |
| 51 | 1.39 | 5.00 | -1.37 | 1.01 |
| 52 | 1.43 | 5.12 | -1.21 | 1.69 |
Graph:
#not centered
dataset4_average %>%
ggplot(aes(week, avg_nap_hrs)) +
geom_line() + geom_point() +
labs(x = "Week", y = "Avg nap time (hrs)",
title = "Average Nap time, by week")
dataset4_average %>%
ggplot(aes(week, avg_awake_hrs)) +
geom_line() + geom_point() +
labs(x = "Week", y = "Avg awake time after nap (hrs)",
title = "Awake time between nap and bedtime, by week")
#centered
graph_combo <- dataset4_average %>%
pivot_longer(c(nap_centered, awake_centered),
names_to = "measure", values_to = "hours") %>%
mutate(measure = recode(measure,
nap_centered = "Nap time",
awake_centered = "Awake time after nap"))
graph_combo|>
ggplot(aes(week, hours, color = measure)) +
geom_line(linewidth = 0.9) + geom_point() +
scale_x_continuous(breaks = seq(0, 52, 4)) +
labs(x = "Week", y = "Average hours", color = NULL,
title = "Average Nap Time and Awake Time After Nap, by Week") +
theme_minimal() +
theme(legend.position = "bottom")
Summary
Over the year, the toddler napped an average of 1.72 hours per day and stayed awake for an average of 4.82 hours between the end of the nap and bedtime. To compare weeks, I standardized both measures as z-scores centered at 0 on these yearly averages, so values above or below 0 mean a week was above or below 1.72 or 4.82 hours.
Over the course of the year, the two measures moved in opposite directions. Early in the year, weekly nap time was above the yearly average and awake time was below it. By the end of the year, nap time was below average and awake time was above it. From week to week, however, the relationship was not consistent. In some weeks, nap time and awake time were both above or both below average. Without a statistical test, it is hard to say whether the two are truly correlated.
#
#