Engaging the community around this fighting game challenged me to rethink my approach to creating data visualizations

Jane Zhang

Follow

Sep 3, 2020 · 17 min read

It’s been a full year since I’ve become an independent data visualization designer. When I first started, projects that came to me didn’t relate to my interests or skills. Over the past eight months, it’s become very clear to me that when clients hire freelancers, they want either: dashboards, interactive visualizations, or reports. I didn’t care for any of these. It always surprises me when I get emails for this kind of work because my portfolio is far from what they look for. I specialize in static visualizations that aim to achieve two outcomes: data documentaries (like a film documentary, except it recounts events through data), and visualizing the world (integrating data into existing formats that don’t traditionally express data, like a food menu).

I considered quitting data visualization altogether in June 2020 and moving onto something else. I originally entered the field because I was inspired by creative work done by Giorgia Lupi, Nadieh Bremer, Stefanie Posavec, Eleanor Lutz, Mona Chalabi, The Pudding, and so many more. I didn’t join this industry to spend the majority of my career creating dry client work. By August, however, I received two new opportunities which both aligned with my interests and skillset — a reminder that despite the outcome, I just need to remain patient during this journey.

I learned that the independent market thrives in operations that can scale: courses, speaking, workshops, and books. Work usually comes from folks in business analytics, research, health, and non-profits. For work that doesn’t scale, like consulting or services, it’s tough to get a sustainable roster of clients who want creative work that goes beyond the standard bar chart.

So, how do I move forward? I looked back to my previous article and thought about the people I talked to. I’ve tried freelancing, teaching, and speaking. The only option left is to develop products, specifically print work for sale. My skills thrive in static data visualization, so naturally, print made sense. I wondered about designing expected things like posters, reference guides, or books. However, I had no experience and was afraid to try these out.

Miraculously, one project idea came to me in the middle of June, it was going to be based on Tekken, a 3D Japanese fighting game that’s been around since the late 90's. I’ve been around games for as long as I can remember. When I was a kid, my older brother would put me right beside him as he played Dreamcast or Playstation 2. Gaming is close to my own identity, so it’s fitting that it’s what would give me another chance in this industry.

See the full project I did on Tekken.

How to Think About Something

I often think that how we think matters more than what we know. I wanted to challenge data visualization and see if it can teach others how to think about something instead of focusing on the what.

The chart from the Financial Times shows the explosive growth of online retail during the pandemic. The approach of this chart is to expose what happened to online retail during the pandemic. The reader can see there has been a sharp growth and the annotations provide context to the story. This approach of showing the what is an important facet of data visualization. But, I was more interested in showing the how.

I arrived at this idea of showing the how after watching a YouTube video describing the combat systems of video games; it opened my eyes to how combat works and the deeper mechanics involved with it. I learned about start-up animation, which is extremely important during combat. This is an animation that occurs as you attack, immediately before your move impacts the enemy. Light attacks are faster than heavy attacks because they have fewer frames in the start-up animation, whereas you risk being more vulnerable and subject to attack from enemies with heavy attacks.

Once I learned the mechanics, I inherited a new perspective on combat. If the mechanics of something are revealed, I could understand the how. Ever tried to take a mechanical pencil apart? I’ve done this when I was younger to get the lead unstuck. After doing this several times, I had a general understanding of how the pencil worked.

How can data visualization teach readers how to think about something? I don’t have a thorough answer as I’m still in the early stages of thinking about this. However, through my own reflection, one way is to compartmentalize information. Like how scientists look to understand the world through the microscopic scale, information is no different.

I first started exploring this idea with a board game called Bang! I broke down how the game works by visualizing the distribution of the cards. A simple histogram revealed which cards were more valuable due to scarcity. I’ve played Bang! over 100 times, but I’ve never taken a moment to think about which cards were more rare. Furthermore, this visualization shows how the game was designed, favouring offensive play.

While The Tekken project was an evolutionary step in my desire to show the how, I never expected it to also completely change my design process.

How Tekken Changed My Process

I stumbled on an amazing video about how Mick Gordon creates music for the latest installments of the Doom games. In his talk, he repeats the following insight:

“If you want to get a different outcome, you must change the process.”

Tekken catalyzed the shift in how I work and there were several things I discovered in the new process.

Multidimensional data

The most seductive aspect of exploring Tekken as a data visualization project was its multidimensionality.

In Tekken 7 Season 3, you can play as one of 30+ characters. Each character has over 100 fighting moves you can use. Generally, players start by having one ‘main’ or the character they will specialize in. This is important because they have to learn and master their character’s most effective moves, and gain a general sense on how to combat the other 30+ characters. It’s extremely difficult to play this game well due to the amount of knowledge and experience it demands.

Players have voluntarily collected data for each move per character. A popular website players reference is rbnorway. In addition, many use Tekken Chicken on mobile to access this data on-the-go.

In general, each move is defined by the following properties:

Command input: The buttons you press on your controller to execute the move.
Hit level: In general, a move can hit three areas on the body. High, middle, and low hits will hit the head, torso, and legs respectively. High and middle hits can be blocked while standing, but low hits require you to block while crouching. In addition, you can dodge high hits by crouching, and you can dodge low hits by jumping. Keep in mind this is a general rule.
Damage: The amount of health you would deplete if your move successfully hits your opponent.
Start-up frames: Every move has an animation before the move hits the opponent. Some moves have shorter animation (i.e. fewer frames). The fewer the frames, the faster the move, the less vulnerable you will be.
Follow-up frames on hit: When your attack hits an opponent who isn’t blocking or in the middle of an animation, you could have plus or minus frames that could help you link moves more effectively. This phenomenon is also known as frame advantage/disadvantage.
Follow-up frames on block: Similar to frames on hit, but this applies when you opponent blocks your move.
Follow-up frames on counter hit: Similar to frames on hit, but this applies to when your move lands when your opponent is in the middle of an animation.

If that was too confusing, here’s a short video I made describing this data. I have added one more variable, effective range. This refers to the maximum distance you can be for the move to still hit.

It’s simply not enough to look at these variables in isolation, when every move has a variety of properties that define it. The data in this game presents the perfect playground for data visualization. There’s no other medium that works better than data visualization to explain these moves in a holistic way. Every video or article I’ve come across would use this data to help explain why it’s effective in certain situations. They present the data as plain text. Though, there are limitations with this because it’s hard to compare these moves for context.

I notice I’m often drawn to datasets that can be visualized as a form of small multiples. Earlier in 2020, I’ve been making experimental projects that can summarize characteristics of something through small multiples:

I was drawn to the idea that the world is multifaceted. There isn’t one thing that describes anything. If someone has a medical condition, to understand it, various lifestyle factors needs to be considered: diet, sleep routine, work, relationships, religion, exercise, childhood upbringing, relationships, trauma, and the list goes on. Reality is complex.

Designing with the Tekken community

While developing this project, I engaged with a public forum for the first time. I accessed Tekken players by joining the community where they hung out: the Tekken subreddit. I wasn’t a regular Reddit user so I had to learn how the platform worked as I engaged with the community.

In total, I engaged with the community four times.

1. The first was to ask for suggestions and check if what I wanted to make already exists. See it here.
2. The second was to validate my prototype and understand if it was worth pursuing. See it here.
3. I reached out to one of the users who had been giving me feedback. I asked them to provide one-on-one feedback because receiving feedback from many users at once was overwhelming and took a toll on me. They gave me feedback on Version 2.
4. I declared the project complete and shared the final version as a public post in order to gauge people’s reactions. See it here.

I was naive to engage with people who weren’t familiar with data visualization this way because the response was brutal. All kinds of feedback were thrown at me and I had to pick which suggestions to filter out. Still, I knew that I was onto something interesting. This chaotic process was simply good user-centered design. In order to eliminate the one-sidedness of things and improve overall accessibility, I listened to user problems, understood how they read data, figured out ways to solve their problems.

Taking feedback from the public like a champ

In designing the first version of this project, I made three mistakes:

1. I made too many assumptions about what my target audience needs. E.g. creating something that helps people build combos, when in reality, learning sample combos provided in the game was sufficient.
2. My target audience was too vague.
3. I made a legend that was too complicated.

My initial goal was to create something that would help beginner to intermediate players better understand the game and help them build tactics and combos to win the game. But, this shifted after creating Version 1.

This visualization shows five moves that Josie Rizal can play. They are part of a list of her most effective moves.

I shared Version 1 with the Tekken Reddit community and received interesting responses. There were a few that particularly struck me.

The feedback below is negative. But, I found it interesting to read. It hit on something that I always think about. Do people in general find complicated data visualizations overwhelming? Their main point was that I was simply visualizing a data table that people can simply reference. Why bother learning how to read a data visualization if a data table is more accessible? These are interesting points and I think it questions how visualizations need to add a lot of value for people to learn how to read it and access insights.

This feedback about clarifying my goals is constructive and positive. If the problem and the goal aren’t clearly defined, the result would feel ambiguous. And, this user used the term ‘data viz’. I never used that term in any of my posts to avoid confusing jargon. It seems to me they have familiarity with it and that’s why their feedback was constructive.

I really liked the honest feedback here. It seems that this user isn’t part of the data visualization or design world, but they did their best describing how I could improve my visualization while tackling the technicalities. It was nice to see them try to break down what worked and what didn’t.

I’ll be honest, it wasn’t easy reading through all the feedback I got. I felt like I missed the mark on something and wasn’t sure if it was worth it to keep pushing forward. After taking a few days to mentally recover, I adjusted my perspective and started to see the silver lining: I was lucky enough to tap into a community that cared enough to give me such detailed feedback. They love Tekken and that’s why they are active on Reddit. Feedback is never easy to accept and unfortunately, it’s much easier to focus on the negative comments rather than the most insightful ones. I was adamant that I wanted to listen and focus on what people were trying to say. I took a deep breath and began working on the next iteration.

Validating a user need

It was clear to me what I had to do. I had to start over and reinvent my process. I didn’t have a clear goal because I didn’t have a clear problem to solve. So, I went to the beginner’s megathread on Tekken and went through the questions new players asked, I needed a way to synthesize them. Somehow, I started recording each question I came across on a sticky note.

I had about fifteen noted and began arranging them on my desk. The next step was to synthesize. I clustered similar sticky notes together and named each cluster based on their shared similarities. I noticed a trend: all the questions varied on a spectrum of specificity. Questions ranged from being very vague (looking for direction on how to start learning) or very specific (need help with technical aspects of the game). I decided to focus on challenges people had related to character-specific tactics.

I finally had a breakthrough. I finally felt like I was walking on even ground. I was so lost the whole time because I had no idea what I was doing. Once I identified a specific problem pertaining to a specific target audience, I felt relief. Now, I could back up claims pertaining to why I was doing this and test every possible solution against its problem.

Design what data visualization does best: the case for physical cards

There are a lot of better ways to tell a good story, such as movies, stand-up comedy, graphic novels, music, vlogs, and so much more. Why do it through data?

Before I tackled this project, I challenge if data visualization makes sense. For example, the video below is a timeline of the history of women’s rights. It could also exist as a large infographic. Whichever is more appropriate would depend on the context.

What makes a static infographic different from a video in this case? The main difference is comparison, a fundamental function that data visualization enables. In this video, each year is shown one at a time. If you made a large infographic, you could compare multiple years at a time to allow for deeper analysis.

When working on the Tekken project, I had to understand why data visualization made more sense as a communication medium over other ones. The most important factor to consider was the sequential nature of how information is communicated.

Let’s break this down.

When it comes to learning about Tekken, players access information through online forums, articles, and YouTube videos. In terms of content, I am competing with articles and videos. Articles and videos are extremely valuable in providing structure. However, there are limitations: the flow of information is sequential and cannot be changed.

A video flows scene by scene. The viewer has no control over how information is presented.

Similarly, an article is structured by paragraphs. One follows the next and the reader has no control over this.

In data visualization, this format is usually seen through scrollytelling visualizations.

To change the nature of information flow so that it’s not sequential, there needs to be a component of interactivity so people can control what they see. This example from The Pudding explores masked wrestlers. There are two ways to experience the data: 1) guided stories and 2) skip the story and explore the data on your own.

The guided stories are sequential and follow a specific order.

Explore mode allows users to filter as they please based on nationality, decade, and theme of the masks. You have freedom to see all the masks.

This is the power of interactivity, it allows for control. In this case, the user can determine what they learn and they can make their own conclusions. This is a healthy mix of both explanatory and exploratory data visualization. Users can follow a pre-determined story and also have a chance to further explore. This is also commonly known as the Martini Glass Structure.

A challenge with a web-based solution in the case of Tekken is that it requires a computer. Players can play Tekken on their computers or on a console. Players who play on their computers will have difficulty referencing data visualization online.

The other option is to create a poster or a printed guide that can be used on-the-go. The challenge here is that information is fixed in place and there’s little control on how it’s presented. The final option I came up with were physical cards. Information is not sequential and it’s still easy to reference while practicing.

What’s interesting about this solution was that the cards are essentially small multiples. Ultimately, the function of these cards is to enable comparison. You can rank them through any variable, create your own categories, and learn how to think about Tekken. I think this format suits this need very well and it adds value in ways that other mediums of communication cannot.

Engaging new communities

I have always been interested in injecting data visualization into the world and letting lay audiences see a different way to view information. Data visualization practitioners are already aware of this potential. The rest of society, not so much. I call what I do a success if I can get people outside of my peer community to see its usefulness. Comedian Ms. Pat describes this really well:

If I can make that crazy looking white boy laugh and that old white man laugh, and that thug black boy laugh, that’s when I know the joke gonna work. Cus’ women gonna support women regardless, when I can go out and get a man to laugh at me going through menopause, to me, that’s funny.

Source: Comedian Ms. Pat Shares Her Life Story, Selling Crack, Getting Shot & More, Breakfast Club Power 105.1 FM.

Unfortunately, I think a conundrum that tends to happen in any creative industry, is that our work tends to be seen by our peers, more so than our target audience. Data visualization is no exception. On Twitter, data visualization people follow me. On Instagram, it’s designers. As creators, we rarely get followed by people who could be potential clients. It’s usually peers who follow us because they are interested in our process and techniques. Here’s an excerpt that perfectly captures this:

…The goal of the visualization was to engage readers in finding and telling their own stories in the data. It was hoped that residents in various occupations would engage in social data analysis, sharing expertise from their respective industries. Despite good intentions, the visualization largely failed in this goal. A total of 23 people submitted 62 comments, with 25 of these comments being posted by the producers of the visualization. Other guests pointed out trends of interest and shared pointers to other related data sets; for example, a registered nurse shared his first-hand experiences in the Health sector. However, the majority of posters were not citizens of Minnesota; they were visualization and statistics enthusiasts drawn by the technology (the piece was mentioned on a popular visualization blog) and not by the story.

Source: Narrative Visualization: Telling Stories with Data by Edward Segel and Jeffrey Heer.

I am thinking more about how to engage relevant audiences to my work. Moving beyond my peer community and going into spaces where I am a nobody. It’s terrifying as heck. But it’s interesting. What made this project worth it at the end was when an experienced Tekken player told me this:

I’ve not seen anyone with your skill set take an interest in Tekken. I’m excited to see the end product. Hopefully, it’ll inspire others to do something similar.

It made my heart skip a beat. It was exactly what I wanted to do. I wanted to explore untapped communities who have never thought that data visualization could do something interesting for them. It’s a small attempt at my vision of visualizing the world.

If you’d like to see more documentation of this project, you can see the full blog post and the YouTube process videos I made.

Jane Zhang is a data visualization designer based in Toronto, Canada. She designs learning experiences through bespoke visualizations. You can find her work at www.janezhang.ca and connect with her on Twitter, LinkedIn, and Instagram.