Ario Barin Ostovary

Software Engineering, University of Waterloo

Sonify

Voice-first analytics that let merchants hear trends instead of reading charts.

Charts assume you can see them. I wanted to know whether a merchant could understand how their store was doing purely by ear, so I built Sonify as an accessibility experiment: a voice-first analytics assistant for Shopify that answers questions out loud and plays the data back as sound.

The core idea is sonification. You press Talk and ask something like "How are we doing today?" The agent replies with a short spoken answer, then optionally plays a clip where pitch encodes the value of a metric and loudness encodes how much it is changing, with distinct earcons marking spikes, dips, and anomalies. There are presets for a single trend, for comparing two ranges, and for layering revenue against orders.

I built two modes so it serves different needs. Explain Mode narrates the data for someone who wants the story, and Listen Mode favors the sonification with short cues like "spike on February 24" for someone who wants to skim by ear. On top of that there is a daily audio briefing, an anomaly listen, and an A/B comparison of two date ranges by sound. The app is React Router 7 with Prisma and a Backboard voice agent, ElevenLabs handles the speech, and I also wrote a standalone Python dashboard that sonifies thirty days of traffic versus revenue just to show the idea quickly.