Cycle Safe Seattle

Background

This project is part of HCDE 511 - Information Visualization, a graduate level course under UW's Human Centered Design and Engineering program. Taught by Professor Holger Kuehnle & PhD teaching assistant Brett Halperin.

A complete data visualization process was carried out over the course of ten weeks. First, we stared with the infrastructure and collisions data set from Seattle Open Data. We then cleaned, analyzed and created visualizations based on the aforementioned dataset. We also incorporated feedback from usability testing, and considerations from user research based on the feedback from our primary and secondary users. 

Outcome

We delivered a complete data visualization dashboard built with Tableau. Through research and usability testing, we validated our target users, confirmed our data story layout, and design for data visualizations. In addition, we generated then implemented design recommendations on types of data displayed, interface, and general improvements to the final data visualization dashboard.

Role

Data Analysis & Visualization
Interaction Designer

Tools

Tableau
Mapbox
Seattle Open Data

Time

10 weeks (Jan - Mar 2022)

Team

Raf Laus,
Jennifer Spriggs,
Sierra Jenkins &
Aichen Sun

A preview of existing bike infrastructure in Seattle.

Data obtained from Seattle Open Data

My Contributions

I took on the responsibilities of obtaining data, preprocessing, analysis, visualizing data, contributing to user research via prototyping, and implemented changes from usability testing. I primarily analyzed quantitative, nominal, ordinal and geospatial data. I also built the dashboard using Tableau, and synthesized various maps from my collaborators. Details of my contributions are listed below.

UX Research & Usability Testing

User research planning, script writing, prototype, and implemented findings from user testing to final data visualizations & dashboard. 

Data Analysis

Cleaned the collisions dataset, exploratory data analysis, investigate data types for encoding / mapping, combined secondary datasets

Information Visualization & Design

Sketched initial concepts, data encoding, created visualizations, animation, data storytelling & final data visualization dashboard using Tableau.

Discover

Data Profile

Seattle Open Data Collisions Dataset from the Seattle Department of Transportation (SDOT) Yearly auto, pedestrian and bike collisions from 2004 to 2021. Over 37 data dimensions from geospatial, nominal, ordinal, and quantitative data. 

Ordinal

severity

Quantitative

date/time
# of vehicles involved
# of injuries

Nominal

collision type
address type
weather
road conditions
light conditions

Geospatial Data

latitude & longitude
polygons for area boundaries (e.g. neighborhoods and cities)
lines (city streets)
points (addresses, incidents)

Define the problem statement.

Problem Statement

How might we use storytelling, and data visualization to provide value to the Seattle cycling community?

Users

Primary & Secondary Personas

1      The Group Bike Lead

Group ride leaders take special considerations, over and above what they would typically plan for themselves, when planning routes for rides involving larger groups of people. These routes tend to be designed with safety in mind while considering the volume of a group.

2     Bicycling Activists

This data can be useful for the involved cyclist activist either individual or involved with an organization like Cascade Cycle Club, who is interested in influencing the city to prioritize building safer bike infrastructure in their communities.

Project Goals

Bike Collision Hotspots

Examine the dangers bikers face in certain conditions by highlighting instances and frequency of collisions and accidents involving bikes across the city.

• How has the total number of bike collisions changed over the years
• How severe are they?
• Where are the dangerous hot spots in the city?

Collision Conditions

Study the existing bike infrastructure and its relationship to biking safety in Seattle — and how certain conditions contribute to dangerous cycling conditions.

• What are the conditions that contribute to bicycle collisions?
• Road conditions?
• Weather conditions?
• Intersection type?

Develop

Initial Sketches

• Where are the crashes present in the city?

• What is the volume per location?

• What years had the most crashes?

• Do collisions spike at certain times of the day?

• Does that change depending on the day?

Tableau visualization experiments

The distribution of bike collisions the day BEFORE, and the day AFTER COVID.

Challenges

1. Integrating & Narrowing Data

Combining our primary and secondary data sets
Narrowing in on the most relevant dimensions of our complex data

2. Combining Geo Data

How much we can layer on a map before it’s incomprehensible, or cumbersome data-wise for the system?

Deliver

Usability Test A

Tableau Story Format

• Wanted to test out early, low-fi versions of the visualizations using the Tableau story format
• Tested comprehensibility of the charts and visualizations with users
• Tested the direction of the concept and desirability of the data for cyclist personas

User Feedback

• Want to see the dangerous intersections contextually situated on a map
• Want to be able to filter the data based on what’s more relevant to them such as a specific neighborhood or type of bike infrastructure
• Wanted to see the full narrative of the data, not just a slideshow
• ⅔ participants mentioned it wouldn’t necessarily alter their own riding behavior, but it would be helpful to them when planning for group rides

STORY & EDITORIAL

"It would help me with friends or group rides that I may go on or lead in the future to avoid those areas if they are specifically very treacherous.

I almost want a map of ghost bikes where people have died so I can zoom in and see markers of where fatalities have happened."

P1, Local Cyclist

Usability Test B

Tableau Editorial Format

• Adjusted page layout to read more as an editorial rather than a slideshow.
• Situated data in map form with neighborhood boundaries visualized
• Included visualization specifically about fatalities to highlight ‘ghost bike’ narrative

User Feedback

• “Having the streets are critical, and the numbers of the arterials.”
• Narrative threw her off
• Wanted the transition to be bike collisions after bike infrastructure
• Wants the highest volume routes — where are people riding? Even though they know we don’t have ‘count data’ everywhere.

“Fatal collisions is something we are highlighting in part because the way you get funding for new infrastructure is based on a scoring system.”

P2, Structural Engineer

Final Deliverable

The final result was a combination of the editorial and story format built with Tableau. The visualization is best viewed on a wide screen. Here are the two versions: Story and Editorial

Extensions and Improvements

• Toggle by the year + time the infrastructure was built that would give you the story.
• IDEAL: What did the network look like in 2015 and what were the crashes in 2015
• Volume data!
  • The highest volume routes — where are people riding? Even though we don’t have ‘count data’ everywhere.
• Responsive / Mobile Layout

Raf Laus, Portfolio. https://rlaus.xyz. © 2024