Kieran Yuen — nonprofit data operations

Nonprofit data operations

Weekly reporting for 26 schools, without anyone logging in

case study · salesforce · apps script · bigquery

An entire leadership layer was pulling its own numbers out of Salesforce, one person at a time. I replaced that with an individualised report that arrives in their inbox every Monday morning — built around the fact that the people who most needed the data were never going to log in to get it.

Schools reporting weekly: 26
Logins required: None

Includes the decisions that mattered — among them throwing out a weighted priority score, because nobody trusts a number they cannot reason about.

Summer melt milestone tracker — a working demo, on invented data, of a tool built for a college access program. Click around it: the controls are the argument.

Built in the open

How late is your NJ Transit train?

live · python · github actions · public data

A scheduled pipeline over NJ Transit's published performance data. It reads eighteen files every week from a source that intermittently fails, merges rather than rebuilds so a failed download can never delete history, and publishes figures that anyone can recompute. The work I do at my job I can only describe. This one you can read.

Counting every late train: 85.5%
Amtrak-adjusted: 92.1%

Northeast Corridor, the twelve months to June 2026. NJ Transit publishes both figures and leads with the second. Every number here is recomputed from their own files by a verification script that fails loudly if any published figure drifts.

Also running

Expense Dashboard — logs a purchase from my phone in two taps, straight into a Google Sheet. Answers one question: am I ahead or behind for the month?
Morning Digest — an email that arrives before I'm awake with the day's tasks, pulled from Notion and grouped by life area.
Activity Log — in development. Records what I actually did today, and whether it gave energy or drained it.

Earlier builds

Through Line — a life-management app that makes accumulated progress visible. React, Supabase, and Postgres, installed on my phone.
Persist — a college search tool over federal College Scorecard data for 2,423 schools, built around one question: where can I afford to go and actually graduate?
Sun Time — my first native iOS app. Tells you how much daylight is left.
Matters — a journaling and goals app organized around life pillars. React, Supabase, Vercel.

About

I've been building things in spreadsheets since I was twenty. The tools got bigger; the instinct didn't. What I keep running into in nonprofit work is the same gap — the data exists, it's even correct, and it still isn't reaching the person who has to make a decision with it. That last stretch is the work I'm best at and the work I find most worth doing.

The side projects are the same instinct at small scale. I notice a bit of friction in my own week, and instead of tolerating it I go build the smallest thing that removes it. Most of them are a Google Sheet, a script, and a trigger. They're still running, which is the only review that counts.

How I work

The tools and methods behind these

Get in touch

hello@kieranyuen.com