| home page | data viz examples | critique by design | final project I | final project II | final project III |
Using my general outline from final project I, looking at pedestrian-related collisions in Allegheny County, I began to become more familiar with Tableau and my extensive data set from PennDot/Western Pennsylvania Regional Data Center. The more I looked at the data, the more difficult it became to decide on what “story” I wanted to tell–or at least, what I felt I could tell without making unfounded claims or presenting information in a misleading way. I eventually decided that the lack of a clear common thread connecting pedestrian collisons could in itself point to a larger issue.
I used Shorthand to begin building out my idea, and was able to fill in most sections except for the final call-to-action area.
The Shorthand project can be viewed through this link: https://carnegiemellon.shorthandstories.com/4838d8bb-87ae-4f2f-a0c3-ce01d0f7c6e0/index.html
My data story is targeted to individuals living in Pittsburgh, and especially residents who have experience moving through the city as pedestrians. Given this—and the fact that I know very few people outside of CMU—I decided that three of my peers at Heinz would be representative of Pittsburgh residents who have experience walking through different neighborhoods and intersections in the city. If I had more time and connections, I would have preferred to interview people covering a wider geographic area and age range to see if they relate to the content differently and could provide other insights based on lived experience.
I kept the script simple, based on the example we were provided in class. The goal of my user research was to understand more generally if I had gained enough skills in Tableau (starting from 0) to create charts that were easy to interpret and made sense with the data. Since I struggled a bit to find a common thread in my dataset after combing through it, I also wanted to know if the different elements of the project seemed to cohesively connect into one larger story.
| Goal | Questions to Ask |
|---|---|
| Understand if visualizations and storytelling themes are effective | What worked? / What didn’t work? |
| See if there are gaps in my logic or in how I presented data | What questions came up? |
| See which storytelling elements stuck with users or engaged them most | What new inspiration arose? |
| Questions | Student 1 (26 years old) | Student 2 (33 years old) | Student 3 (25 years old) |
|---|---|---|---|
| What worked? | The use of colored text to highlight important words was effective, and the initial bar graph is very clear and simple. Listing the streets was great and it was an engaging way to tell a story. The “Most Concentrated High-Risk Intersections Were in Downtown Pittsburgh” figure was also done really well and could be a figure I can see myself using | I really like the personal connection to the Linden/Wilkins intersection. I liked highlighting how many collisions happened at traffic lights. It really fits with my experience of Pittsburgh drivers being crazy at yellow lights. I liked highlighting specific intersections but being able to see the map with other areas. I learned about different dangerous intersections in neighborhoods I’m not normally in. | I liked the breakdown & analysis of how these collisions happened around traffic signals. Street view for the crash statistics was really helpful too. |
| What didn’t work? | The “Most Intersection Collision Sites” image felt too long in its current placement and was difficult to read the surrounding text due to the length. It looks like it can be shortened by removing some white space. | Question under a question - I wasn’t sure which question the answer was for. The text boxes that are over images are a little too transparent to read as clearly. Maybe make the background more opaque? For the part that talks about data from 2010, I would have appreciated having that graph go back to 2010 instead of just 2020. I found the wording regarding intersection collision site traffic controls kind of confusing.The scroll function gets a little weird on the map. Not sure what to do about it but sort of breaks the habit of scrolling down for more info. – | The Point Breeze death story – What 5-way intersection are you talking about? Would you say its a pretty central/busy thoroughfare? […] Considering the death rate in PGH, why was this the story you wanted the audience to know about? |
| What questions came up? | Im curious about how the high-risk intersection graphic is calculated, specifically, whether the count is per person. Also the colors were nice but I wonder whether the color choices are accessible for colorblind viewers and take into account light text on light background | Is Pittsburgh a vision zero city? What is Allegheny county doing right that it has lower pedestrian deaths? Is it just that so much of Allegheny county is suburban or rural compared to other cities that boost the average? Are there busy intersections that are particularly safe and what are they doing right? | Why are there such concentrations in certain areas? I was thinking maybe because of the amount of people? But that theory doesn’t hold up great… Maybe it could be the roads being main roads where people like to go fast AND that there are so many people in those areas? Much to think about… |
| What new inspriation arose? | This sparked my interest in incorporating dynamic, interactive graphs! Im also curious if this would look good on all platforms (laptops, phones etc) | I feel very fired up about this topic (even more so than before) | I am inspired to never leave my home again. |
User interviews were very helpful in understanding what elements were most effective, and which could use additional work. Below I have broken down some of this feedback by theme:
Images: Multiple users mentioned that they appreciated the Shorthand section showing Google Streetview images of high-risk intersections in Pittsburgh: “ Listing the streets was great and it was an engaging way to tell a story.—Student 1, 26 years old”
Personal connection: I received mixed feedback on including my personal connection to the topic, with Student 2 (33 years old) saying they “really like the personal connection to the Linden/Wilkins intersection,” while Student 3 (25 years old) suggested adding more details and connecting the story more clearly to the rest of the content: “What 5-way intersection are you talking about? Would you say its a pretty central/busy thoroughfare? […] Considering the death rate in PGH, why was this the story you wanted the audience to know about?”
Map: Users cited the map (“Most Concentrated High-Risk Intersections Were in Downtown Pittsburgh”) as helpful, saying it “could be a figure I can see myself using” (Student 1) and that they “learned about different dangerous intersections in neighborhoods I’m not normally in” (Student 3). However, Student 2 also noted that Shorthand’s scroll feature didn’t seem to work well with this embedded visualization. Stacked Bar Chart: One user noted that they liked how this visualization highlights “how many collisions happened at traffic lights” because it matched their “experience of Pittsburgh drivers” (Student 2). However, Student 1 described how the chart could be improved visually, saying the chart “felt too long in its current placement and was difficult to read the surrounding text due to the length.”
Text & Accessibility: Student 1 asked “whether the color choices are accessible for colorblind viewers and take into account light text on light background,” and Student 2 similarly brought up that “the text boxes that are over images are a little too transparent to read as clearly.”
I was pleased to hear that users found the map-based sections particularly engaging or helpful, because I felt these areas were the most important for grounding the data in reality and making the relevance clear to Pittsburgh residents. That said, I have quite a bit of work to do in other areas based on other feedback I received.
First, I’ll be revisiting some of the color elements of the project to make sure that it’s readable for all users. I’ve also gathered that there are areas in both the body text and visualizations that I could clarify what data is being presented and tie it more clearly to my overarching data story. I also need to work within Tableau and Shorthand to better embed the interactive visualizations into the final project. Finally, I need to complete my final informational and call-to-action section of the project, in addition to adding citations throughout.
“Cumulative Crash Data,” Western Pennsylvania Regional Data Center, https://data.wprdc.org/dataset/allegheny-county-crash-data/resource/2c13021f-74a9-4289-a1e5-fe0472c89881
“Pedestrian and Overall Road Traffic Crash Deaths — United States and 27 Other High-Income Countries, 2013–2022,” Centers for Disesase Control and Prevention, https://www.cdc.gov/mmwr/volumes/74/wr/mm7408a2.htm#F1_down
“Searching for the ‘Smoking Gun’ in US Pedestrian Deaths,” Bloomberg, https://www.bloomberg.com/news/articles/2026-04-14/what-s-really-driving-the-pedestrian-safety-crisis-in-us-cities?srnd=phx-citylab