Montreal BIXI bike data

1. Montreal BIXI bike data

Imagine that you are the lead data scientist of the Montreal BIXI Bike sharing company. You have just been given data about every time a bike was ridden in 2017, with over 4 million rides! It is your job to explore this data to start to gain insights into patterns of usage and ultimately help your company make sure routes and stations are adequately stocked to meet predicted demand. This is a good example of a large dataset to which all of the methods we have learned about in this course can be applied. First, aggregations and summary plots can help us get the big picture, and then more detailed displays using Trelliscope can help us dive a little deeper. In this case study, we will carry out several exploratory visualizations that will guide us to insights that we should take into account when starting to build statistical models.

2. BIXI bikeshare data

The BIXI bike data system records every ride's start and end location, time of day, and duration for the year 2017. For the exercises, we have randomly subsampled the data down to 1 million records, but we will make the full 4 million ride dataset available.

3. Let's dive in!

We will start to explore this data using some of the summary visualization techniques that we learned in chapter 1.

Create Your Free Account

or

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.

This exercise is part of the course

Visualizing Big Data with Trelliscope in R

BeginnerSkill Level

4.8+

Start Course for Free

Learn different strategies for plotting big data using ggplot2, including calculating and plotting summary statistics, various techniques to deal with overplotting, and principles of small multiples with faceting, which leads into Trelliscope.

Exercise 1: Visualizing summaries Exercise 2: Daily ride counts Exercise 3: Distribution of cab fare amount Exercise 4: Distribution of payment type Exercise 5: Adding more detail to summaries Exercise 6: Relationship between trip duration and total fare Exercise 7: Faceting daily rides Exercise 8: Tip amount distribution faceted by payment type Exercise 9: Visualizing subsets Exercise 10: Comparing fare distribution by payment type Exercise 11: Visualizing all subsets

In the previous chapter you saw how faceting can be used as a powerful technique for visualizing a lot of data that can be naturally partitioned in some meaningful way. Now, using the trelliscopejs package with ggplot2, you will learn how to create faceted visualizations when the number of partitions in the data becomes too large to effectively view in a single screen.

Exercise 1: Faceting with TrelliscopeJS Exercise 2: Trelliscope faceting gapminder by country Exercise 3: Interacting with the TrelliscopeJS displays Exercise 4: Interacting with the display Exercise 5: Additional TrelliscopeJS features Exercise 6: Customizing the gapminder display Exercise 7: Examining the new cognostics Exercise 8: Adding your own cognostics Exercise 9: Adding custom cognostics Exercise 10: Interpreting custom cognostics

The ggplot2 + trelliscopejs interface is easy to use, but trelliscopejs also provides a faceted plotting mechanism that gives you much more flexibility in what plotting system you use and how to specify cognostics. You will learn all about that in this chapter!

Exercise 1: Trelliscope in the tidyverse Exercise 2: Grouping and nesting Exercise 3: Stock price display Exercise 4: Exploring the display Exercise 5: Cognostics Exercise 6: Adding cognostics Exercise 7: Cognostics from nested data frames Exercise 8: Navigating stock plots with new cognostics Exercise 9: Trelliscope options Exercise 10: Customizing the stock display Exercise 11: Visualizing databases of images Exercise 12: Visualizing Pokemon Exercise 13: The most powerful Pokemon

<a href="https://www.bixi.com" target="_blank">The Montreal BIXI bike network</a> provides open data for every bike ride, including the date, time, duration, and start and end stations of the ride. In this chapter, you will analyze data from over 4 million bike rides in 2017, going between 546 stations. There are many interesting exploratory questions to ask from this data and you will create exploratory visualizations ranging from summary statistics to detailed Trelliscope visualizations that will give you interesting insight into the data.

Exercise 1: Montreal BIXI bike data

Current Exercise

Exercise 2: Number of daily rides Exercise 3: Examining time-of-day Exercise 4: Effect of membership and weekday Exercise 5: Summary visualization recap Exercise 6: Daily plots Exercise 7: Looking at all days Exercise 8: Top 100 routes dataset Exercise 9: Augmenting the data: Route summary statistics Exercise 10: Visualizing the data: Counts by hour-of-day Exercise 11: Evaluating the visualization Exercise 12: Au revoir