AirStory

Version 0.0 was a medium-fidelity wireframe thrown together to prove the data pipeline worked. PM2.5, CO, Temp, Humidity, trend charts, a Daily Reflection prompt — everything in, no UX thinking applied. No empty states, no design system, no user flow. The only question was: does the data show up?

It did. Ship it and move on.

Version 1 introduced group-level organization — students belong to a school, class, period, and group. The dashboard now shows weekly trends and a monthly calendar view, letting students track change over time rather than just a single snapshot.

Station naming (MTN12 · Group 4), data download, and persistent group identity made this the first version that felt like a real classroom tool.

Version 2 is the current deployed system at haetalkim.github.io/tamguingAIR. Built with React, Express, and PostgreSQL, connected to OpenAQ for city-level reference data and Google Maps API for geospatial visualization.

Four main views: Heat Map (real-time air quality across Manhattan, color-coded by AQI), Raw Data (filterable by school/class/period/group, CSV export), Analysis (statistical dashboard with 7-day and 30-day trends), and My Page (student profile, group members, account settings).

Teacher-generated join codes let students register to a class. Role-based access separates instructor and student views.

Accessibility was built in from the start — color coding throughout the platform follows guidelines for visual accessibility, and the Heat Map includes a colorblind-friendly mode that can be toggled on. Data shouldn't be readable only by some students.

Before moving to the next iteration, we ran a structured UX audit of the current platform — evaluating 4 screens and 3 operational guides from both teacher and student perspectives. The goal: find what's broken before real students use it.

4 High severity issues: No direct data input for students (CSV only — unrealistic for K-12 field collection). No empty state on first login — blank charts look broken. Session creation buried — teachers can't find it on day one. Numbers shown without AQI context — PM2.5 = 18 means nothing without a "Good / Moderate / Unhealthy" badge.

3 Medium issues: Raw Data defaults to all groups. Join Code UI ambiguous. Heat Map MODE toggle unlabeled.

The audit also mapped the platform against the Instructor Curriculum Guide as a "living pair" — where disconnected, both needed to update together. These findings are now feeding directly into the next build cycle.

Version 3 is the direct response to the UX audit. Almost everything flagged as high priority got addressed: proper login flow with Student / Teacher portal separation, Manage Classes for teachers (Join Code generation, class structure setup), Analysis now shows your data against OpenAQ reference location — so students can actually compare their measurements to a real baseline. Raw Data now has real field measurements in it (Morningside Park, March 2026).

The platform also now properly surfaces the three-tier comparison in Heat Map: City Average, School Average, and Group Average side by side — the core of what makes SSI inquiry possible.

The mobile app is currently in development. Once complete, students will connect directly to the sensor via the app — replacing the current CSV upload flow with real-time wireless data transmission. No laptop needed in the field.

A school launch is planned for Summer 2027 with a contracted partner school. The curriculum guide (Student Guide + Teacher Guide) is written and iterated. The hardware is ready. The platform is live.