Robi's Game






3 Months




Primary school students may feel it hard to understand some complex programming knowledge.
Design an educational game that involves how a specific knowledge point (sentiment analysis) is applied to game challenges, and guide students step by step to solve the problem, to increase their interests and help them better understand this learning content.
Provide a high-level flow and supporting wireframes.


Teach students to learn the structure and application of "sentiment analysis" through a in-class mini-game.

Sentiment analysis” is a natural language processing (NLP) technique used to determine whether data is positive, negative or neutral. Students might find it hard to know how this technique is applied for actual problem solving, and the process structure behind this technique.


About the Background.

This mini-game is supplementary material to the course content. In the class time, after learning the basics, students will experience the game to learn more about how the machine performs “sentiment analysis” and how it can be used to solve a real-world problem.

First, let's Target and focus on a specific sentiment analysis method as game content.

I chose textual analysis as the main focus, because:
-collect and detect text information is simple and requires less workload to build a game
-keep interactions easy and consistant for students

Then, analyze and draw the User flow of the Text Sentiment Analysis.

Here is main flow for textual sentiment analysis, I extracted each key step. And identify learning points (marked in red) for students. They are design challenges for the build of a easy-to-play eduactional game.


So what are some insights I have ?

(1) It’s important to give a context of this game quickly: including a story, problems and game goals.
Students find it hard to connect with abstract algorithm knowledges, and important concepts are usually forgotten quickly when they don’t point to specific experience or scenarios.

(2) Constructing a “Dictionary/Lexicon” to analyze emotions will be the main game challenge.
How to design the game flows, and the feedback loop to make students understand the importance and structure of the emotion dictionary?

(3) Negative words and degree adverbs help understand the complexity of textual sentiment analysis.
How to introduce and apply these two concepts in the game?

(4) Some less important process could be simplified(like disposing meaningless words).
In the actual problems, meaningless words are controlled by other dictionaries and disposed automatically by AI, this process will confuse students if put too much emphasis on it.

Robi's Game

Now, Let Me Introduce the Game: Robi's Restaurant

Robi’s restaurant is one of a series of educational games to teach students about basic AI knowledges. It’s designed to help students understand the main structure and application of textual sentiment analysis, which is a common way for AI to analyze emotions of comments, posts and other text-base contents.

01. How will students experience this game?

Start with introducing an interesting story in the restaurant, students will train the little robot to help his uncle read restaurant comments, and judge them are negative or positive.

02. How to introduce game rules to students?

Students will learn the rules from dialogues with little Robi at first, he helps students to select and identify negative and positive words through yellow dialog boxes below, and guide students on how to play the game.

03. How to Understand the complexity of textual sentiment analysis?

Students will be asked to select three parts of speech, collect and score words by themselves based on words’ emotions. Students will learn the structure and  application of  emotion dictionaries through this practice.

04. How to encourage students and evaluate the game result?

By launching the dictionaries, students will see if their work help Robi read comments successfully. There is a final share screen to show the accuracy of the game result, students also have a chance to improve their dictionaries to get a perfect result!


My Role

This is a team project about creating a series of educational games to teach students about basic AI knowledges. In this mini-game, I collaborated with  programmers and visual designers to build this fully-functional web game. I provided text contents, game strategies, low-fi and high-fi wireframes, also supported several iterations with this game.

My design goals

When I made sure of the form and scale of the game. I need to reframe the design goal ,by clarifying the users needs, players personas and environment settings.


first, I did researches on teaching content- spent several hours figuring out differences and structures of main kinds of sentimental analysis, and found several daily scenarios to apply this teaching point. After that, I talked with our content strategist, who was familiar with different education levels, and had clear ideas about teaching goals. At this time, I also asked the tech support to know better about limitations about our teaching devices.

What did I learn from this part?


I did competitive analysis on several educational games. I compared a few aspects including game narratives, interactions, teaching strategies and analyzed advantages and disadvantages.

What did I learn from this part?


Once I had enough information about what the content and game experience would be like, I started to ideate on storytelling and game challenges. My goal here is to arrange several points to an easy-to-understand game solution. I chose the most common usage for textual analysis – about customer service, and arranged the game narrative around it.

Design Decisions

After ideations and several discussions, I made decisions on several design details.

Why I create the story about Robi helping review comments to explain the sentimental analysis?

Why I design the process of select different kinds of dictionaries?

Why I provide the button of auto-score emotional words?

Why I design the process of iterate the dictionaries?

Why I made the share page to show the successful rates?

Iterations & Sketches

THanks for reading!