Introduction

Pokémon is anime and gaming franchise that was started on the 2nd of February in 1996, and since it’s inception the franchise has been unstoppable with over 120 games, almost 1200 anime episodes, and many other features for it toys, manga, and trading cards portions. The basic premise of Pokémon takes place in world where there are creatures called Pokémon that have amazing abilities, and where humans team up with Pokémon to train against and battle others. For my final project I took a small portion of this world to make an interaction dashboard in Tableau Public; additionally, I conduction user research to help build and refine my project.

Overall, I started this project with the following goals:

1. Learn more about how users interact with the Pokémon franchise
2. Design an informative and interactive dashboard that would be engaging to Pokémon fans
3. Teach those not as familiar with Pokémon about some basic properties of the world of Pokémon

Inspiration

There are multiple sources inspiration behind this final project, the first being my own love for the franchise. Since about the age of 6, I’ve been an avid Pokémon fan having watched a lot of the original series, and played most of the main games for the Nintendo systems. So, this was definitely a project I was rather excited and knowledgeable about. Within the Pokémon franchise there is a device known as a Pokédex which indexes and provides information about all Pokémon known to the series, I decided to make my own Pokédex that could be used for someone looking to learn more about the various battle stats associated with every Pokémon.

One other source of inspiration would be my network graph from lab 3. In this lab I explored visualizing different top anime organized in a network by how related each one was to one another. This lab as well as the feedback on my network led me to think more about anime, Pokémon, and data visualization.

Materials

User Research

To conduct my user research, I decided to do semistructured interviews and monitored users while browsing the Tableau dashboard. I recruited 2 participants for the my research that were members of a PokémonGo group I discovered through mutual friends. We were able to meet on Zoom so I could remotely monitor the participants’ experience with my final project. During the Zoom call, I took notes based on areas where I noticed someone struggling and feedback I received during and after the session.

Data and Visualization

My final project uses two different data sets both found on Kaggle. One features all the stat information for each Pokémon and can be found here. The other is used as the source for the pictures in the Pokémon graph on the first dashboard and was found here. To organize and structure my data I used Microsoft Excel, and, lastly, to piece everything together and create my final project I used Tableau Public.

Methodology

User Research

Utilizing a semistructured interview format, I was able to guide my user research session while also allowing more input from my participants. I conducted both of my user interviews during a mid fidelity version of my interactive dashboard. I started building my dashboard based on my own assumptions with only a few stats and interactive elements, but before adding I wanted to get user feedback to make sure I was on the right track. My user interviews consisted of pre-study questions before viewing Tableau, discussion during viewing, and post-study questions. I used the following format to conduct my study:

• Pre-study questions:
  • How old are you?
  • Approximately how long have you been a fan of the Pokémon franchise
  • How knowledgeable would you say you are about Pokémon? Beginner, Intermediate, Expert?
  • Are you familiar with battle stats? What do you know about them?
  • When you think of a specific Pokémon, what do you want to know about them?
  • What’s your favorite Pokémon game and Pokémon?

After answering those questions, participants then proceeded to explore the Tableau dashboard. I encouraged participants to first explore and then vocalize what they thought they were doing and what different sections of the visualization meant to them. I made sure to make note of sections where participants appeared to be confused or spent the most time on. During this portion and after participants were done with their initial exploration, I asked the following questions:

• Post-study questions:
  • What is the overall feeling you get from the different tabs / dashboards?
  • Do you find it useful or not? Why?
  • What is your favorite tab / dashboard? Why?
  • What’s missing?
  • What was unclear?

After asking the questions, I went through and talked a bit more about my plan for the visualization and demonstrated any functionality that the participant might not have known about during their exploration. And then asked if they had any final comments or recommendations.

Data and Visualization

Set up:

To build my project in Tableau, I first had to make sure both of my data files were formatted the same way. The only common column between the two datasets was the “Pokémon name” field. I made sure to format them the same way and then began to upload them into Tableau Public. I utilized the JOIN feature in Tableau based on the Pokémon names and decided to do an INNER JOIN to exclude any null values. This, however, did remove some Pokémon with more complicated names that weren’t standardized across both datasets, but most of the main and non-special Pokémon were unaffected by this inner join.

I knew I wanted this visualization to include images so to set that up I had to make sure that my data and my image files were arranged alphabetically. Then, I moved my Pokémon image folder into the “shapes” folder within my Tableau folder on my computer. These steps enabled to match each Pokémon to the appropriate image in Tableau.

Tableau:

Since I knew I was working with Pokémon, I knew that the type of each Pokémon would be main field I’d be working with. I did face a problem because some Pokémon have more than one type, so I decided just to display main types for this visualization. I first made the treemap visualization and segmented the color based on Pokémon type. I matched each type with the appropriate color that is used in the games. Next, I made the shape graph with the images of each Pokémon. Both the treemap and shape graph were sorted by the attack battle stat initially. I then made three other treemaps that featured the three main types of Pokémon that every Pokémon game starts with: Fire, Water, and Grass. I wanted to include a side-by-side comparison of the three. Lastly, I placed the main treemap and the shape graph on a dashboard and used the treemap as a filter for the graph so that anything a user selects in the tree map will show the image of that Pokémon on the shape graph. It was at this stage that I conducted my research.

Results

User Research

I will now discuss the results from my user research. One of my users was 21 and the other was 24, both have been playing Pokémon for over 10 years and claimed to Intermediate in terms of Pokémon knowledge, with one participant saying they’re almost an expert. Both were familiar with battle stats and knew most of the different stats (one forgot about the speed stat). Both participants stated they would want to know what a Pokémon looked like, its type(s), description, and at least it’s HP and at least attack or special attack. One participant said would want to know all the stats. In terms of favorites, one of my participants favorite Pokémon was Pikachu, the franchise mascot, and favorite game was Pokémon HeartGold for the Nintendo DS. The other participant was a fan of Charizard and Pokémon Super Mystery Dungeon for Nintendo 3DS.

While viewing my initial version of my final project, one participant had no experience with Tableau, so it was clear they were a bit disoriented and first clicked on the graph, wondering why the images were so small, but then they click on one of the tiles in the treemap and was surprised by the way the dashboard changed. They only clicked on one tile at a time during their exploration of the main dashboard. For the other tabs, they definitely spent less time browsing than the first, main dashboard which was expected. Overall this participant seemed to enjoy the project.

My second participant didn’t seem as intimidated by the Tableau interface. But similarly to the first participant, most of the time was spent selecting random parts of the treemap and watching the Pokémon appear on the graph. This participant thought the other treemaps were really interesting but wanted to see individual treemaps for other Pokémon types.

I noticed the following general trends from my user research:

• Majority of time spent on first dashboard
• More interactive elements are needed throughout my project
• More battle stats should be shown
• The Pokémon chart by itself didn’t seem as useful to participants
• Unsure if the three individual starter type treemaps are necessary
• The styling could be better and more engaging
• Background on the topic might be helpful
• Users had a difficult time discovering the available actions
• Users wanted to see dual Pokémon types

Data and Visualization

Based on this feedback, I updated by visualization to have the following features. I added colors to make the visualization look like a traditional Pokédex to hopefully familiarize the user. I also added some instructions through out the visualization to make sure everyone knows the main features of each dashboard. Next I added a lot of customization through filters on each dashboard, including a search filter so everyone can check and see the stats of their favorite Pokémon and compare them side by side. I made the main treemap filterable by Pokémon type so that a user can see how each type stacks up against one another. I also added a tool tip for all of my treemaps that features all the stats of a particular Pokémon and even a small preview of what that Pokémon looks like. I added the image to the tool tip by creating a sheet that just references my image database which was set up earlier and the name of the Pokémon again. Then, I removed all the formatting, including the names to leave only the images. Lastly, I inserted the this image sheet into the tooltip and it dynamically pulled the corresponding image for each Pokémon. There is also a tool tip for the image chart that features the specific battle stat currently being filtered. I moved the type legend to the bottom of the graph because I thought it was best out the way but still adjacent to treemap.

Reflection

Over the course of this semester, I learned a lot and am proud of how far I’ve come. For this specific visualization, one major challenge I wasn’t able to figure out was how to display Pokémon with more than one type. My work around was to only show the primary type of each Pokémon, this did somewhat skew my visualization. Specifically, the “flying” type Pokémon only consists of two entries. I had a hard time wrapping my head around the best way to present this, but ended up leaving it out. Another thing I would fix is the tool tip feature on the treemap. While I was really happy I was able to figure it out, most, if not all of the images are cut off, so I debated removing this entirely. Additionally, I could not figure out how to use the same filters for my treemap and image graph, so I had to make two separate filters for each. And, the same applies for the starter type Pokémon sheet. I thought it would be confusing to have the same filter on one dashboard three times, so I left them on their individual dashboards and had it dynamically update the treemaps on the combined dashboard. Lastly, I think I could’ve provided a little more context for those not familiar with Pokémon. I think a dashboard with just text explain how the battle stats work in the games would have been a great addition to this visualization. Overall, I’m pretty happy with what I was able to do, I think I really made a lot of essential changes based on the feedback from my user research. I plan on working on the challenges I faced and adding on to this project during this summer.