Personal Project

Scalable Mobile Designs - FE Datamine

The Problem of Scale

Data mining is, by definition, an enormous undertaking. It involves reverse-engineering text, images, audio, and logic, then redisplaying everything found in a flat, consumable format. In my case, we were spending our free time trying to make a 200-hour game with branching paths available in 12 languages discoverable for gamers who didn’t have time to play through the game dozens of times. My work partner and I were the initial main clients of this effort, both as fans of the game and as a development team who wanted to keep their skills sharp.

Step one was clear given this problem - understand what data was available to us. That meant a lot of slogging through spreadsheets and categorizing the chunks we could make sense of. This culminated in an initial site that was more debugging than design to wrap our heads around the scope.

Designing for Internationalization

Early on, we decided that if the game we were mining had been released with twelve different languages supported, we would need to support the same on our site to help our fellow fans. Unsurprisingly, the two of us working on the project didn’t know all those languages. We realized that using Google Translate wouldn’t help us here; if we wrote anything in English, browsers could handle a basic translation themselves without us needing to maintain languages we didn’t understand.

This lead to some creative problem-solving where we used snippets of the game menus to drive much of our navigation. That allowed us to leverage the work already done by the localizers of the game. It required rethinking much of our information architecture since we wanted to minimize the use of hardcoded English outside of our project description. We also had to rely on our own rough knowledge of 3-4 of the supported languages, including the original Japanese, to ensure we didn’t latch onto the wrong titles due to ambiguous translation. Some friendly translators in the community were of great help with this step.

One example of what we faced: “Menu” as in a navigation element vs. “Menu” as in something you order food off of. The game had both, and that meant we initially had a navigation which read ‘food menu’ instead of ‘menu’ in some non-English languages until we realized our mistake.

Audience Drives Scope

As the months dragged on with our dev-heavy work of parsing the data, it became clear we needed to tighten our scope if we intended to go public with the site. Normally, we would have started with this step, but it wasn’t until we asked the gaming community for help in refining our data that we realized just how many other people this could benefit beyond us or translators. Thus began a new definition of who we wanted to serve and what we considered important for them.

Part of this was a thorough branding exercise. The other was an assessment of our community, both from observation and polls. We concluded that we were after three major audiences: casual fans who had played the game and wanted to learn more about their favorite characters without dropping even more hours to explore all the options, translators who wanted to compare between the original language and the localization familiar to them, and fans who enjoyed deep analysis of the game for the enjoyment of the community or writing of fanfiction, often while on their phones.

This made us realize we needed two additional things to release: a solid search across the thousands of lines of game text and a mobile site with no functionality loss.

Single-Functionality Pages

We worked through many iterations for mobile before we realized the key to a successful design with such a robust data set. Only one idea per page. One major interaction to worry about made the pages feel comfortable to our target users, even if a lot of scrolling was involved. Searching and jumping to sections where possible also helped with this.

Unfortunately, meant we ended up with many more pages than we originally anticipated. Every item, every book, every quote category, every battle, and every scene has its own page. It was enormously difficult to test given our two-man operation, even with the help of some peers, and we still got corrections a year later. However, we have never received complaints about the usability of the mobile site since adopting this technique.