INTRODUCTION
The Carnegie Mellon University Graduate Student Association came to the class with a hope to understand the reasoning behind the under-utilization of the CMU transit option, especially the shuttle and escort services. GSA believes that in order to better guarantee the safety of students during commute, changes must be made. The team took on the project by conducting in-depth user research and provide a potential solution with a preferred future that could be easily implemented.
This project lasted 12 weeks from August - December 2019.
MY ROLE
User Researcher
UX Designer
Visual Designer
SKILLS
Contextual Inquiry
Modeling
Think-Aloud
Speed Dating
Log Data Analysis
Classification
Wizard of Oz
TEAM
Gillis Bernard
Stu Mcgibbon
Zach Peng
Melissa Shi
Allana Wooley
Siqi Wu
Overall Methods and Final Insights
00
Research Overview
01
PROBLEMS
The core issue facing Carnegie Mellon University students when it comes to transportation is a difficulty finding information and, if they do find the information, knowing how to use it to get around.
Decentralized transportation information
No personalization on delivered information
Poorly publicized transportation options
02
METHODS
Generative
-
Contextual Inquiry and Affinity Diagrams
-
Background Research
-
Speed Dating and Storyboards
-
Think Alouds
-
Surveys
Evaluative
-
Wizard of Oz Testing
-
Think Alouds
-
5-Second Tests
03
INSIGHTS
From out user research, we identified 3 major insights revealing why the transportation service is currently not reaching to the students body.
Optimization and Centralization: Transit information should be located in a single resource that does the work of comparing options for the user, telling them in one-stop what their best options are
Lazy Learning: Students prefer social learning because of the low effort required to gather information. If they have a friend in the know, all they have to do is ask to get exactly the answer they were looking for.
How-To Instructions: Instructions for using the various transit options are not clear or easy to locate, causing frustration and hesitation from both current and potential student riders.
“I want to learn about my options, I just don’t want to do any of the work myself.”
Background Research
Understanding the Domain
01
When first given this research project, I decided to first educate myself on the domain by conducting experiential research and generate sequence and flow models based on given interview materials. From this, I was able to identify several high level insights and potential future research directions.
Experiential Research
CMU has a partnership with Ride Systems, a digital platform that hosts bus service information. On Ride Systems, the CMU students can access shuttle and escort tracking services which helps them find the right transit potion navigating around Pittsburgh and to and back from CMU. In order to understand the system, I experienced using the app first hand.
Experiential Research Finding 01
Unclear limitation of app functionality creates confusion for users.
Experiential Research Finding 02
Limited information on the app prohibits users from smoothly utilizing the app.
Experiential Research Finding 03
Lack of direct feedback generates huge confusion and frustration for users when setting alarms on the app.
Understand the Current Operation
With the notes from prior interviews that has been conducted with the CMU police lieutenants and shuttle and escort drivers, I generated flow and sequence models to identify the current breakdowns in the system and further research directions.
Lack of Real-time Notification for Students
When the bus is full and the bus driver has to skip bus stops, students do not receive updates about bus stop skipping, leaving them frustrated when the bus does not pick them up.
The Current Transit Service is Not Efficient
The Police department believes that they more funding is necessary to improve the escort and shuttle services utilization, however the services are not fully utilized.
02
Discovering the Underlying Problems
Generative Research
Identify Key Stakeholders
Based on our initial background research and secondary research, we mapped out the major players in the transit system. From there, we selected 4 key stakeholders to conduct contextual interview from which bettered our understanding of the major breakdowns in the system.
Police Lieutenant
GSA
Driver
Student
Interview Stakeholders and Affinitizing
To understand how each stakeholder view and understand the transit system in place, we conducted contextual interviews will them and later affinitize the different perspectives to find high level patterns that are otherwise overlooked.
Rider motivation differs
Stakeholder expectation and reality do not align
Lack of communication channel between rider and service provider
Diverse perspective between stakeholders
Narrowing Research Direction
"CMU students should have a nearly complete mental model of all transportation options available to them through the university, including basic information for use, facilitating the ability to contrast and compare options and efficiently make the best decisions for their needs."
Interview Students Utilizing Different Transportation Methods
We identified 5 main forms of transportation to and from school for CMU students and conducted contextual inquiry with each group of students. After affinitizing all the interview findings, we were able to identify several key insights.
Pittsburgh Bus Rider
Bike Rider
CMU Shuttle Rider
Walk
CMU Escort Rider
No experience is uniformly bad
Learning from trial and error
While all of our participants expressed complaints or pain points at some point during our interviews, many also acknowledged areas where the transit system was working for them.
This prolonged learning period requires time and effort to master and may limit a users willingness to try out alternative modes of transportation, anticipating the same learning curve.
Gather knowledge through social connections
Though CMU provides official information for students to learn and understand the offered transportation services. Many of the students do not use these channels. Rather, they learn about their options through the words of mouth.
03
Validate the Identified Needs for Making Design Decisions
Evaluative Research
Validating the Needs of Students
From the generative research conducted and the insights gathered, the team identified 3 major types of needs and decided to validate them with students through evaluative research - storyboarding. Furthermore, we wanted to test and see the acceptance boundaries of potential solutions by providing several scenarios for feedback/
The need to identify the most optimized transit methods for a user’s particular situation.
The need to learn about and internalize transit options through social and experiential methods.
The need to compare all available transportation options.
04
Identified Design Opportunities
Problems and Insights
Conclusion of System Problems and Insights
After both generative and evaluative methods the team were able to identify 3 main problems that existed in the current CMU transportation system.
Decentralized transportation information
No personalization on delivered information
Poorly publicized transportation options
From out user research, we identified 3 major insights revealing why the transportation service is currently not reaching to the students body.
Optimization and Centralization: Transit information should be located in a single resource that does the work of comparing options for the user, telling them in one-stop what their best options are
Lazy Learning: Students prefer social learning because of the low effort required to gather information. If they have a friend in the know, all they have to do is ask to get exactly the answer they were looking for.
How-To Instructions: Instructions for using the various transit options are not clear or easy to locate, causing frustration and hesitation from both current and potential student riders.
Why CUI?
From research, we learned that students prefer the form of word-of-mouth for gathering information. Students also highly prefer easy access to information, that is why we decided to provide a centralized platform from which they can receive personalized information on which transportation works the best for them. Knowing that Facebook can host a chat bot, and is a commonly used social media platform among students, the team decided to host this CUI on Facebook.
Chat flow version 1
Intent, classification, slot extraction, and prompt version 1
Conversation Flow Testing Through Think-Alouds + Wizard of Oz
To see how well our conversation design matches with the mental model of that of students, the team split up and conducted think-alouds with students using wizard of oz. One team member pretended to be the chat bot and responds to testing participants' messages.
Options were not provided completely in the beginning or after user restart making the chat bot confusing to use.
Graphical representations had low readability.
Semantics used in the chat bot caused a lot of confusion for users because it did not match their expectations.
Refine Conversation Content
Based on the think-alouds feedback, the team made a a series of changes to the conversation content including reducing text length, altering card and map representation, and renaming features.
Information Efficiency Testing through 5 Sec Test
To see how well the information are being delivered, we decided to conduct 5 second testing with students to see what to they remember from reading the conversations and the graphics. Additionally, we tested what their overall impression of the chat bot is.
Placement and color contrast greatly impact element visibility.
Icon is not sufficient to communicate transportation mode.
Graphic contents are more eye catching and memorable.
05
CUI as a Solution
Final Design
Personalized Transportation Option at the Tips of Your Fingers
To provide a central hub that contains all necessary information for a student to painlessly and efficiently identify the right transportation method upon entering CMU, a chat bot is designed to be hosted on Facebook for student to engage in conversations that leads them to necessary information.
Final chat flow version 2
Concise text with visual break points
To avoid text wall, visual breaks are necessary. Images and emojis are used as visual break points to "light up" some of the necessary text walls.
Cards containing different transportation options + advantages
Clear color coding for different transportation options and shuttle routes. On the center top of each cards contains information regarding the superlatives of each transit option.
Clear visual map showcasing selected transportation option
Minimalistic route maps that highlights the start and end points with simple road map of the surroundings.