9 Bike New York City

Dany Rosete

9.1 Introduction - CitiBike in NYC

9.2 CitiBike Data

9.2.1 Description

The CitiBike dataset contains a detailed record of bike trips within NYC. Each row represents a completed ride. With this dataset, it is useful to analyze urban mobility as it includes geographical, time-based and user type data.

Here is the link to the dataset.

9.2.1.1 Structure of the Dataset

This dataset is divided by Month directories that contain multiple .csv files. The complete dataset includes data from 2013 to March 2026 (as of May 3rd, 2026).

Each row corresponds to a single bike trip that has the following fields:

Unique trip identifier (ride_id)
Type of trip (rideable_type): electric vs classic bike
Start time (started_at)
End time (ended_at)
Start station Name and identifier (start_station_name & start_station_id)
End station name and identifier (end_station_name & end_station_id)
Start Longitude & Latitude (start_lng & start_lat)
End Longitude & Latitude (end_lng & end_lat)
Type of Member (member_casual)

This dataset includes the following disclaimer: “The data has been processed to remove (…) any trips that were below 60 seconds in length (potentially false starts or users trying to re-dock a bike to ensure it’s secure).”

9.2.1.2 Data Manipulation

For this project, not all data will be used. Also, since .csv files are 100+ MB (exceeding the 100MB limit from Github), some files will be converted into Parquet format.

To use Parquet files, arrow library is needed. An R script to convert the .csv files to parquet files is found in the /scripts folder.

9.2.2 Missing value analysis

library(arrow)
library(ggplot2)
library(dplyr)
library(lubridate)
library(scales)
library(forcats)
library(redav)

9.2.3 How Weather/Seasons affect use?

citibike_ds <- open_dataset("data/citibike_parquet")

Error:
! IOError: Cannot list directory 'data/citibike_parquet'. Detail: [errno 2] No such file or directory

seasonal <- citibike_ds |>
  mutate(
    hour = hour(started_at),
    wday = wday(started_at),
    month = month(started_at),
    date = as_date(started_at)) |> 
  count(month, wday, hour,date)|> 
  collect() |> 
  mutate(
    season = case_when(
      month %in% c(12,1,2) ~"Winter",
      month %in% c(3,4,5) ~ "Spring",
      month %in% c(6,7,8) ~ "Summer",
      month %in% c(9,10,11)~ "Fall"), 
    season = factor(season, levels = c("Winter", "Spring","Summer", "Fall")),
    wday = factor(wday, levels = 1:7,labels = c("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"))) |>
  group_by(season, wday, hour) |>
  summarise(avg_rides = mean(n), .groups = "drop")

Error:
! object 'citibike_ds' not found

ggplot(seasonal, aes(x = hour, y = fct_rev(wday), fill = avg_rides)) +
  geom_tile(color = "white", linewidth = 0.1) +
  facet_wrap(~season) +
  scale_fill_viridis_c(option = "inferno") +
  scale_x_continuous(breaks= seq(0, 23, 6),
                     labels= paste0(seq(0, 23, 6),":00"),
                     expand = c(0, 0)) +
  labs(
    title = "Citi Bike rides by season (2025-2026)",
    x = "Hour of Day",
    y = NULL,
    fill = "Avgerage Rides") +
  theme_minimal(base_size = 12) +
  theme( legend.position  = "bottom", legend.key.width = unit(1.5, "cm"))

Error:
! object 'seasonal' not found