DATA 606 Data Project Proposal

Author

Patricio Romero

Published

October 21, 2026

Data Preparation

The first step is to load the required packages and import the NYCHA electricity dataset.

Code

library(readr)
library(dplyr)
library(tidyr)
library(ggplot2)
library(knitr)
library(scales)

nycha <- read_csv(
  "Data/Electric_Consumption_And_Cost_(2010_-_Sep_2025)_20261018.csv",
  show_col_types = FALSE
)

The dataset is imported into R and stored in an object named nycha.

Research question

How is electricity consumption associated with electricity cost among NYCHA housing developments in New York City, and how does this relationship vary across boroughs?

This question is limited to the NYCHA billing records included in the dataset. The results cannot be generalized to all buildings in New York City.

Cases

The cases are monthly electricity billing records for NYCHA housing developments. Each row represents an electricity record for a development during a specific billing period.

Code

nycha_cases <- nycha |>
  transmute(
    development_name = `Development Name`,
    borough = Borough,
    revenue_month = `Revenue Month`,
    consumption_kwh = parse_number(
      as.character(`Consumption (KWH)`)
    ),
    current_charges = parse_number(
      as.character(`Current Charges`)
    )
  )

cases_summary <- nycha_cases |>
  summarise(
    total_cases = n(),
    total_developments = n_distinct(development_name),
    total_boroughs = n_distinct(
      borough[
        borough %in% c(
          "BRONX",
          "BROOKLYN",
          "MANHATTAN",
          "QUEENS",
          "STATEN ISLAND"
        )
      ]
    )
  ) |>
  transmute(
    `Total Cases` = comma(
      total_cases,
      accuracy = 1
    ),
    `Total Developments` = comma(
      total_developments,
      accuracy = 1
    ),
    `Total Boroughs` = comma(
      total_boroughs,
      accuracy = 1
    )
  )

kable(
  cases_summary,
  caption = "Summary of NYCHA Cases"
)

Summary of NYCHA Cases
Total Cases	Total Developments	Total Boroughs
553,666	381	5

The output shows the total number of billing records, developments, and boroughs represented in the dataset.

Data collection

The data were collected from administrative electricity billing records for New York City Housing Authority developments. The records contain electricity consumption and cost information associated with developments, boroughs, and billing periods.

Type of study

This is an observational study. The researcher did not assign electricity usage, electricity cost, borough, or billing period. Therefore, the analysis can describe associations but cannot establish causation.

Data Source

The data were obtained from NYC Open Data.

Dataset: Electric Consumption and Cost
Publisher: New York City Housing Authority
Source: NYC Open Data — Electric Consumption and Cost

The original file is stored in:

Data/Electric_Consumption_And_Cost_(2010_-_Sep_2025)_20261018.csv

Describe your variables

The following chunk selects and prepares the variables used in the analysis.

Code

nycha_clean <- nycha |>
  transmute(
    development_name = `Development Name`,
    borough = Borough,
    revenue_month = `Revenue Month`,
    consumption_kwh = parse_number(as.character(`Consumption (KWH)`)),
    current_charges = parse_number(as.character(`Current Charges`)),
    kwh_charges = parse_number(as.character(`KWH Charges`))
  ) |>
  mutate(borough = toupper(trimws(borough)))

valid_boroughs <- c(
  "BRONX",
  "BROOKLYN",
  "MANHATTAN",
  "QUEENS",
  "STATEN ISLAND"
)

glimpse(nycha_clean)

Rows: 553,666
Columns: 6
$ development_name <chr> "SHEEPSHEAD BAY", "SHEEPSHEAD BAY", "SHEEPSHEAD BAY",…
$ borough          <chr> "BROOKLYN", "BROOKLYN", "BROOKLYN", "BROOKLYN", "BROO…
$ revenue_month    <chr> "2025-09", "2025-09", "2025-09", "2025-09", "2025-09"…
$ consumption_kwh  <dbl> 33001, 36068, 30969, 40235, 33956, 26999, 41264, 3057…
$ current_charges  <dbl> 5106.45, 5581.02, 4791.96, 6225.81, 5254.21, 4177.71,…
$ kwh_charges      <dbl> 1523.76, 1665.38, 1429.94, 1857.78, 1567.86, 1246.63,…

development_name and borough are qualitative variables. revenue_month is a time variable. consumption_kwh, current_charges, and kwh_charges are quantitative variables.

The main outcome variable is current_charges, and the main comparison variable is consumption_kwh. Borough will be used to compare the records across locations. The scatter plot includes a visual linear trend line to help describe the direction of the association. The trend line is used for visualization and is not presented as an inferential regression model. If a regression model were used, current_charges would be the dependent variable.

Relevant summary statistics

The following chunk calculates summary statistics for electricity consumption and electricity cost.

Code

consumption_statistics <- nycha_clean |>
  summarise(
    observations = n(),
    mean_consumption = mean(consumption_kwh, na.rm = TRUE) / 1000,
    median_consumption = median(consumption_kwh, na.rm = TRUE) / 1000,
    sd_consumption = sd(consumption_kwh, na.rm = TRUE) / 1000,
    minimum_consumption = min(consumption_kwh, na.rm = TRUE) / 1000,
    maximum_consumption = max(consumption_kwh, na.rm = TRUE) / 1000
  )

kable(
  consumption_statistics,
  digits = 2,
  col.names = c(
    "Observations",
    "Mean Consumption (MWh)",
    "Median Consumption (MWh)",
    "Standard Deviation (MWh)",
    "Minimum Consumption (MWh)",
    "Maximum Consumption (MWh)"
  ),
  format.args = list(big.mark = ","),
  caption = "Summary Statistics for Electricity Consumption"
)

Summary Statistics for Electricity Consumption
Observations	Mean Consumption (MWh)	Median Consumption (MWh)	Standard Deviation (MWh)	Minimum Consumption (MWh)	Maximum Consumption (MWh)
553,666	30.22	11.44	48.26	0	2,788.77

Code

cost_statistics <- nycha_clean |>
  summarise(
    observations = n(),
    mean_cost = mean(current_charges, na.rm = TRUE),
    median_cost = median(current_charges, na.rm = TRUE),
    sd_cost = sd(current_charges, na.rm = TRUE),
    minimum_cost = min(current_charges, na.rm = TRUE),
    maximum_cost = max(current_charges, na.rm = TRUE)
  )

kable(
  cost_statistics |>
    mutate(
      observations = comma(
        observations,
        accuracy = 1
      ),
      across(
        -observations,
        ~ dollar(.x, accuracy = 0.01)
      )
    ),
  col.names = c(
    "Observations",
    "Mean Cost (USD)",
    "Median Cost (USD)",
    "Standard Deviation (USD)",
    "Minimum Cost (USD)",
    "Maximum Cost (USD)"
  ),
  caption = "Summary Statistics for Electricity Cost"
)

Summary Statistics for Electricity Cost
Observations	Mean Cost (USD)	Median Cost (USD)	Standard Deviation (USD)	Minimum Cost (USD)	Maximum Cost (USD)
553,666	$4,339.03	$2,259.86	$6,302.43	-$25,622.21	$424,298.64

The first table reports the number of observations, mean, median, standard deviation, minimum, and maximum for electricity consumption. The second table reports the same statistics for electricity cost. Separating the variables makes the results easier to read and compare.

Summary statistics by borough

Code

borough_summary <- nycha_clean |>
  filter(
    borough %in% valid_boroughs,
    !is.na(borough),
    !is.na(consumption_kwh),
    !is.na(current_charges)
  ) |>
  group_by(borough) |>
  summarise(
    observations = n(),
    average_consumption = mean(consumption_kwh, na.rm = TRUE) / 1000,
    median_consumption = median(consumption_kwh, na.rm = TRUE) / 1000,
    average_cost = mean(current_charges, na.rm = TRUE),
    median_cost = median(current_charges, na.rm = TRUE),
    .groups = "drop"
  ) |>
  arrange(desc(average_consumption))

kable(
  borough_summary |>
    mutate(
      across(
        c(average_cost, median_cost),
        ~ dollar(.x, accuracy = 0.01)
      )
    ),
  digits = 2,
  format.args = list(big.mark = ","),
  col.names = c(
    "Borough",
    "Observations",
    "Average Consumption (MWh)",
    "Median Consumption (MWh)",
    "Average Cost (USD)",
    "Median Cost (USD)"
  ),
  caption = "Electricity Consumption and Cost by Borough"
)

Electricity Consumption and Cost by Borough
Borough	Observations	Average Consumption (MWh)	Median Consumption (MWh)	Average Cost (USD)	Median Cost (USD)
STATEN ISLAND	4,143	107.75	63.20	$14,132.55	$12,689.84
BRONX	125,657	34.23	12.60	$4,917.68	$2,540.16
MANHATTAN	141,415	33.70	7.60	$4,884.52	$2,441.42
BROOKLYN	212,402	25.56	11.76	$3,613.67	$2,100.87
QUEENS	64,276	23.70	14.96	$3,539.93	$1,867.22

This table compares the typical electricity consumption and electricity cost among the boroughs.

Frequency of qualitative variables

The following table reports the number of records for each borough. This summarizes the qualitative variable used in the analysis.

Code

borough_frequency <- nycha_clean |>
  filter(borough %in% valid_boroughs) |>
  count(borough, sort = TRUE) |>
  transmute(
    Borough = borough,
    Observations = comma(n, accuracy = 1)
  )

kable(
  borough_frequency,
  caption = "Number of Records by Borough"
)

Number of Records by Borough
Borough	Observations
BROOKLYN	212,402
MANHATTAN	141,415
BRONX	125,657
QUEENS	64,276
STATEN ISLAND	4,143

This table shows how many electricity records are available for each borough.

Visualization: Electricity consumption and cost

Code

ggplot(
  nycha_clean |>
    filter(
      borough %in% valid_boroughs,
      !is.na(consumption_kwh),
      !is.na(current_charges)
    ),
  aes(
    x = consumption_kwh / 1000,
    y = current_charges,
    color = borough
  )
) +
  geom_point(alpha = 0.4) +
  geom_smooth(
    method = "lm",
    se = FALSE
  ) +
  scale_x_continuous(
    trans = pseudo_log_trans(sigma = 10),
    breaks = c(0, 500, 1000, 1500, 2000, 2500, 3000),
    labels = comma,
    name = "Electricity consumption (MWh)"
  ) +
  scale_y_continuous(
    labels = dollar_format(
      accuracy = 1,
      scale = 1
    )
  ) +
  coord_cartesian(xlim = c(0, 3000)) +
  labs(
    title = "Electricity Consumption and Cost by Borough",
    x = "Electricity consumption (MWh)",
    y = "Current charges (USD)",
    color = "Borough"
  ) +
  theme_minimal() +
  theme(
    axis.text.x = element_text(
      angle = 45,
      hjust = 1,
      vjust = 1
    )
  )

Each point represents one electricity billing record. The plot examines whether higher electricity consumption tends to be associated with higher electricity cost.

Visualization: Electricity cost by borough

Code

ggplot(
  nycha_clean |>
    filter(
      borough %in% valid_boroughs,
      !is.na(borough),
      !is.na(current_charges)
    ),
  aes(x = borough, y = current_charges, fill = borough)
) +
  geom_boxplot(alpha = 0.7, outlier.alpha = 0.25) +
  scale_y_continuous(
    trans = pseudo_log_trans(sigma = 1000),
    breaks = c(
      -25000,
      -10000,
      -5000,
      0,
      1000,
      5000,
      10000,
      50000,
      100000,
      200000,
      350000
    ),
    labels = dollar_format(
      accuracy = 1,
      scale = 1
    )
  ) +
  labs(
    title = "Distribution of Electricity Charges by Borough",
    x = "Borough",
    y = "Current charges (USD)"
  ) +
  coord_flip() +
  theme_minimal() +
  theme(legend.position = "none")

This boxplot compares the median, variability, and possible outliers of electricity charges across boroughs.