cyclistic_data_analysis_project.Rmd

---
title: "Cyclistic - Data analysis project"
author: "Daniel R. Beckert"
date: "2024-01-19"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Objective
Document the analysis process performed for the Google Data Analytics Professional Certificate capstone project.


### Setup
To run this project you need to run the following commands if you don't have `tidyverse`, `skimr` or `scales` installed. 
```
install.packages('tidyverse')
install.packages('skimr')
install.packages('scales')
```

```{r echo=FALSE}
Sys.setlocale("LC_ALL", "English")
```

### Load necessary information
```{r}
library('tidyverse')
library('skimr')
library('scales')
```

## Load datasets
Load all the data necessary to perform the analysis, we are assuming the data is on the `trips` directory on the same folder as this file
```{r}
trips_2022_11 <- read_csv('./trips/2022/11/202211-divvy-tripdata.csv')
trips_2022_12 <- read_csv('./trips/2022/12/202212-divvy-tripdata.csv')
trips_2023_01 <- read_csv('./trips/2023/01/202301-divvy-tripdata.csv')
trips_2023_02 <- read_csv('./trips/2023/02/202302-divvy-tripdata.csv')
trips_2023_03 <- read_csv('./trips/2023/03/202303-divvy-tripdata.csv')
trips_2023_04 <- read_csv('./trips/2023/04/202304-divvy-tripdata.csv')
trips_2023_05 <- read_csv('./trips/2023/05/202305-divvy-tripdata.csv')
trips_2023_06 <- read_csv('./trips/2023/06/202306-divvy-tripdata.csv')
trips_2023_07 <- read_csv('./trips/2023/07/202307-divvy-tripdata.csv')
trips_2023_08 <- read_csv('./trips/2023/08/202308-divvy-tripdata.csv')
trips_2023_09 <- read_csv('./trips/2023/09/202309-divvy-tripdata.csv')
trips_2023_10 <- read_csv('./trips/2023/10/202310-divvy-tripdata.csv')
```

## Prepare and process the data 
```{r}
skim(trips_2022_11)
```

To perform the a standard cleaning process we will create a function. The cleaning process will remove columns that won't be used during the analysis process, remove rows with missing values and fix problems where `ended_at` is before `started_at`
```{r}
clean_dataset <- function(dataset) {
  
  # Remove unnecessary columns
  ds_removed_columns <- select(
    dataset, 
    -c(
      start_station_name, 
      start_station_id, 
      end_station_name, 
      end_station_id,
      start_lat,
      start_lng,
      end_lat,
      end_lng
    )
  )

  # Remove missing values 
  ds_removed_missing <- drop_na(ds_removed_columns)
  
  # Fix problem with dates
  ds_fixed_dates <- ds_removed_missing %>% 
    mutate(
      new_started_at=if_else(started_at > ended_at, ended_at, started_at),
      new_ended_at=if_else(started_at > ended_at, started_at, ended_at),
      started_at=new_started_at,
      ended_at=new_ended_at
      ) %>% 
    select(-new_started_at, -new_ended_at)
  
  return(ds_fixed_dates)
}

```

Create a single dataframe with all the data from the previous 12 files
```{r}
all_trips <- rbind(
  trips_2022_11,
  trips_2022_12,
  trips_2023_01,
  trips_2023_02,
  trips_2023_03,
  trips_2023_04,
  trips_2023_05,
  trips_2023_06,
  trips_2023_07,
  trips_2023_08,
  trips_2023_09,
  trips_2023_10,
  deparse.level = 0
)
```

Clean all the data
```{r}
all_trips_cleaned <- clean_dataset(all_trips)
```

Enrich our data computing the amount of minutes spent in each ride and also the day of the week
```{r}
all_trips_cleaned_enriched <- mutate(all_trips_cleaned, trip_in_minutes=as.numeric(difftime(ended_at, started_at, units="mins")))
all_trips_cleaned_enriched <- mutate(all_trips_cleaned_enriched, weekday=wday(started_at))
```

Remove outliers. Outliers are all those entries above the threshold of 97.5% of rides
```{r}
upper_bound <- quantile(all_trips_cleaned_enriched$trip_in_minutes, 0.975)

all_trips_filtered <- all_trips_cleaned_enriched %>% 
  filter(trip_in_minutes <= upper_bound)
```

## Analyze the data
Identify how much each group represent on the total amount of rides
```{r}
trips_count_df<- all_trips_filtered %>% 
  group_by(member_casual) %>% 
  count() %>% 
  ungroup() %>% 
  mutate(percentage=`n` / sum(`n`)) %>% 
  mutate(labels= scales::percent(percentage)) %>% 
  rename(count=n)
```

```{r echo=FALSE}
ggplot(trips_count_df, aes(x="", y=percentage, fill=member_casual)) + 
  geom_col(color="black") + 
  geom_text(aes(label=labels), position=position_stack(vjust=0.5)) +
  coord_polar(theta="y") +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Contribution of rides per customer type") + 
  labs(x="", y="")
```

Identify how much each group represents in terms of time spent during rides
```{r}
trips_time_df<- all_trips_filtered %>% 
  group_by(member_casual) %>% 
  summarize(sum=sum(trip_in_minutes)) %>% 
  ungroup() %>% 
  mutate(percent=`sum` / sum(`sum`)) %>% 
  mutate(labels= scales::percent(percent))
```

```{r echo=FALSE}
ggplot(trips_time_df, aes(x="", y=percent, fill=member_casual)) + 
  geom_col(color="black") + 
  geom_text(aes(label=labels), position=position_stack(vjust=0.5)) +
  coord_polar(theta="y") +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Contribution of time spent in rides per customer type") + 
  labs(x="", y="")

```

Check the average of time spent in each ride for each group
```{r}
average <- all_trips_filtered %>% 
  group_by(member_casual) %>% 
  summarize(avg=mean(trip_in_minutes))
```

```{r}
average
```

Analyze the distribution of rides throughout the year
```{r}
ggplot(all_trips_filtered, aes(x=month(started_at, label=TRUE), fill=member_casual)) + geom_bar(position="dodge") +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Distribution of rides throught the year") + 
  labs(x="Month", y="Rides count")
```

Analyze the distribution of rides throughout the weekdays
```{r}
ggplot(all_trips_filtered, aes(x=wday(started_at, label=TRUE), fill=member_casual)) +
  geom_bar(position="dodge") +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Distribution of rides throught weekdays") + 
  labs(x="Weekday", y="Rides count")
```

Analyze the distribution of rides throughout a day
```{r}
ggplot(all_trips_filtered, aes(x=hour(started_at), fill=member_casual)) +
  geom_bar(position="dodge") +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Distribution of rides throught a day") + 
  labs(x="Ride started at (hour)", y="Rides count")
```

Analyze the distribution of rides in different days of the week
```{r}
all_trips_weekday <- all_trips_filtered %>%  
  filter(weekday!=1 & weekday!=7)

all_trips_weekend <- all_trips_filtered %>%  
  filter(weekday==1 | weekday==7)
```

```{r echo=FALSE}
ggplot(all_trips_weekday, aes(x=hour(started_at), fill=member_casual)) +
  geom_bar(position="dodge") +
  facet_grid(~wday(started_at, label=TRUE)) +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Distribution of rides through the week (excluding weekends)") + 
  labs(x="Ride started at (hour)", y="Rides count")
```

```{r echo=FALSE}
ggplot(all_trips_weekend, aes(x=hour(started_at), fill=member_casual)) +
  geom_bar(position="dodge") +
  facet_grid(~wday(started_at, label=TRUE)) +
  guides(fill=guide_legend(title="Customer Type")) +
  ggtitle("Distribution of rides through weekends") + 
  labs(x="Ride started at (hour)", y="Rides count")
  
```