Assignment 3

Goals

Use different visualization tools, including Tableau, Observable Plot, and Data-Driven Documents (D3), to create a stacked bar chart.

Instructions

You will need to complete this assignment in multiple parts. All the results can be inserted into a single Observable Notebook, but you may submit a web page containing the results as well. You must use Tableau Public (or the Desktop version), Observable Plot, and D3 for this assignment. All D3 visualization should be done using the 7.x version. You may use other libraries (e.g. lodash) as long as they are not used to construct or modify SVGs, but you must credit them in the HTML file you turn in. Tableau offers video tutorials, and Observable Plot has documentation and examples For D3, there is extensive documentation available as well as examples, and Vadim Ogievetsky’s example-based introduction and the bar chart examples that we went through in class are also useful references.

Due Date

The assignment is due at 11:59pm on Wednesday, October 11.

Submission

You should submit any files required for this assignment on Blackboard. If you use Observable, submit the .tar.gz or .tgz file that is generated from the export menu and rename it to a3.tar.gz or a3.tgz. If you create your own files, please make sure the filename of the main HTML document is a3.html. Any other files should be linked to the main HTML document accordingly relatively. Blackboard may complain about the files; if so, please zip the files and submit the zip file instead.

Details

In this assignment, we will be working with data from the City of Chicago’s Traffic Crashes Data. This dataset details traffic crashes in the city over the past few years. Rather than using this dataset directly, I have extracted the 2022 data as a subset that is available as a csv file. Each record has a CRASH_DATE property storing the full date and time of the crash along with its FIRST_CRASH_TYPE which indicates the type of crash. We are interested in analyzing the number of types of crashes per month. The data is available here:

https://gist.githubusercontent.com/dakoop/78efac022b7fa30cab8bf261285c432a/raw/f70dedfa4ab43f8b429056efc63da8a2efded82a/chicago-traffic-crashes-2022.csv

We will be using Tableau (Public), Observable Plot, and D3 to create a vertically stacked bar chart. The visualizations should show each month along the x-axis and a bar with the total number of crashes each month. That bar should be split into subunits for each type of crash, where each crash type receives its own color. Order the bar fragments appropriately. Provide a legend describing which colors map to which result.

As with Assignment 1, make sure the beginning of your main web page (or notebook) contains the following text:

• Your name
• Your student id
• The course title (“Data Visualization (CS 627/490)”), and
• The assignment title (“Assignment 3”)

If you used any additional JavaScript libraries, please append a note indicating their usage to the text above (e.g. “I used the jQuery library to write callback functions.”) Include links to the libraries used. You do not need to adhere to any particular style for this text, but I would suggest using headings to separate the sections of the assignment.

1. Tableau (25 points)

Either use Tableau Public or download Tableau Desktop and register here to receive a free academic license (You may work before your license arrives using a 14-day trial). Load the CSV file and create a stacked bar chart. If you are using Tableau Public, make sure the publish your visualization (required to save it), and put a link and image in your notebook/web page. If you are using Tableau Desktop and wish to turn in the workbook (.twb file), you may also do that, but put an image in your notebook/web page and a note about the twb file.

Hints

• There is documentation on creating a stacked bar chart available.
• Tableau provides options for the legend ordering and color assignment if you click on the legend (menu in upper-right corner)
• You can choose to group data by month and count entries
• Think about which colors to use. We will discuss this design choice more later in the course.

2. Observable Plot (30 points)

Observable notebooks automatically import the d3 and Plot libraries so you can use them directly. However, if you are using another tool for this assignment, make sure to include these two libraries. See the documentation about how to include them, specifically about how to use them in vanilla HTML or as a UMD bundle. To load the data, use d3.csv which can load the data from the URL above. Note that converting the data types after reading the csv or using D3’s autoType functionality when loading will help avoid parseInt calls and Date constructors later on. You do not need to match the colors in the example solution, but you should make sure that your colors are in the same order for each month. Check the margins and make sure you have a legend.

Hints

• Consider the bin transform to group data by month.
• Investigate the sort option to order the individual bars in a stack.
• Examine the options for the x-axis, including formatting and rotating the tick labels
• Look at the layout options for margins
• There are also many options for the legend

3. D3 (30/45 points)

In this part of the assignment, use D3 to create the same (stacked) bar chart. CS 490 students need only create a bar chart showing the totals per month with appropriate axes and labels, but CS 627 students should create the stacked version with a legend. For both parts, we will need the data to be organized differently (by month). You can use the following code which assumes your original data is stored in crashes:

crashesMonthType = d3.rollup(crashes, v => v.length, d => d.CRASH_DATE.getMonth(), d => d.FIRST_CRASH_TYPE);
crashesForStack = [...crashesMonthType.entries()].map(d => ({...Object.fromEntries([...d[1].entries()]), "Total": d3.sum(d[1].values()), "Date": d[0]}))

This groups the crashes by a numeric month (January is 0, December is 11), and then creates a new array with the month (Date), the total number of crashes (Total), and entries for each crash type where the key is the crash type, and the value is the number of crashes of that type. CS 490 students can then access the values from the Total attribute.

For CS 627 students, it may be easier to first create a non-stacked bar chart (use the provided method above), and then try the stacked version. The axes and labels should be similar to Part 2. The stacked version should have a legend that indicates the relationship between the bar components and the colors. This legend can be created using the Swatches code in this notebook. In Observable, you can do

import {Swatches} from "@d3/color-legend"

in one cell and then pass your color scale (color) to Swatches, appending to another d3 selection divElt as follows:

const swatches = Swatches(color);
divElt.append(() => swatches)

Note that you’ll want to create a div that holds the SVG and the swatches if you take this approach.

Hints

• d3.scaleBand is useful for bar charts.
• D3 has routines (e.g. d3.axisLeft) to build an axis given a scale
• A group element with a transform can help shift the entire visualization so that labels or axes have space (margin convention)
• To obtain the maximum monthly total, you can use d3.max with a proper accessor
• d3.stack may be useful here. Check how the keys function works with the stack.
• If you use d3.stack, you will need nested selections (note that these groups may be different than you expect)