Lens
Cross-Cultural Trust Translator
A mobile-first web tool that translates complex "Urban/Scientific" product information into "Rural/Trust-based" visual languages, directly addressing the accessibility gaps I observed during my Jiaxian internship.
Challenge
Design for "Non-WEIRD" (Western, Educated, Industrialized, Rich, Democratic) populations, translating technical information into culturally appropriate visual formats.
Outcome
A working prototype that maps scientific nutrition data to locally meaningful visual representations, addressing the trust deficit in standard labeling.
My Role
Research, Design, Python Development
Origin: The Ms. Zhang Moment
During my internship at Jiaxian Food Technology, I met Ms. Zhang, a 58-year-old rural consumer who distrusted our yogurt's nutrition label. "The numbers mean nothing to me," she said. "I trust what my neighbors tell me." This moment revealed that standard information design excludes entire populations.
Research Foundation
I conducted 30 surveys and 20 interviews with rural Chinese consumers. Key findings: numbers alone don't convey meaning; trust comes from relatable sources; visual comparison (not absolute values) drives understanding.
The Translation Algorithm
Users select a product category and input standard data (e.g., "Sugar: 12g"). The algorithm maps these to local mental models, instead of showing "12g Sugar," the screen displays 3 Sugar Cubes or a color-coded Health Scale. A "Neighbor Verified" badge addresses the trust deficit.
Technical Implementation
Built with Streamlit (Python) for rapid UI prototyping. The core is a JSON mapping structure translating "Scientific Terms" to "Rural Dialect/Icons." This architecture allows for cultural customization.
Reflection
"My internship taught me that standard information design excludes rural populations. I didn't want to just write a report about it; I wanted to build a solution. I built Lens to prove that accessibility is a systemic problem, not just a design problem."
Key Learnings
- 1Universal design often reflects WEIRD assumptions
- 2Trust is a design element that must be intentionally built
- 3Local mental models are more powerful than "objective" data