I Was a GIS Analyst a Few Weeks Ago. Here's What Changed.

Every career starts at 0°N, 0°E. The interesting part is picking a heading.

First, an apology. The Dispatch went quiet for a couple of weeks — the first gap since this newsletter started. I'd love to claim it was a deliberate sabbatical. It wasn't. The reason is this issue's Main Event, and once you read it, I think you'll understand why the writing slipped. The weekly cadence resumes now.

The Signal

GeoAI Just Got Its Own Conference. That Means Something.

Earlier this month in Ghent, Belgium, the discipline crossed a quiet threshold: the 1st International Conference on Geospatial Artificial Intelligence — GeoAI 2026 — held its inaugural edition, organized by Ghent University's GeoAI Research Center. Pre-conference workshops ran June 3, the main conference June 4–6, with best paper awards, travel grants for early-career researchers, and sponsorship from HERE, Esri, and the major remote sensing journals.

If you've been reading this newsletter for a while, you know the pattern. GeoAI work has lived as a track inside other people's conferences: a symposium at AAG, a workshop at SIGSPATIAL, a themed track at GEOINT. When a research community graduates from "track at someone else's event" to "event," that's not just logistics. It's the field declaring it has enough distinct questions, methods, and people to stand on its own.

What it doesn't mean is that the prototype-to-production gap closed. The hard problem we've covered all spring — getting from a model that performs on a benchmark to a workflow that runs on Tuesday with the data you actually have — is still the hard problem. A first conference is where a field organizes its questions, not where it answers them. The proceedings worth reading will be the ones grounded in operational constraints rather than leaderboard scores.

Chris's Take: I'm glad this exists, and I'm watching it with the same posture I bring to every foundation model paper: show me the workflow. The most useful thing GeoAI can do over its next few editions is build a culture where "we deployed this and here's what broke" earns more prestige than "we beat the benchmark by 1.2 points." The field that figures that out first wins the practitioners.

→ geoaiconference.org

The Main Event

From Analyst to Supervisor: A Field Report from the First Weeks

A few weeks ago, my job title changed. I went from GIS analyst to GIS supervisor. Same desk, same coffee mug, completely different job. This is the field report I wish someone had handed me — written for everyone earlier on the same road, because most of what I'd read about this transition was either LinkedIn platitudes or management theory written by people who've never had to explain a versioned geodatabase to anyone.

The first thing nobody tells you: the promotion isn't a reward for being good at GIS. It's a bet that you can make other people good at GIS. That sounds like a poster in an HR office until you live it. My output used to be measurable — layers, scripts, dashboards, runbooks. Things that rendered. Now my output is whether the analysts on my team are unblocked, developing, and pointed at the right problems. The map I'm responsible for doesn't render anywhere. You make peace with that quickly, or you spend your first months sneaking back into the geoprocessing pane to feel productive — quietly stealing the exact work your team needs to own. I've caught myself doing it twice already. I expect I'm not done catching myself.

The second thing: the calendar is the job now. Introductory meetings with leadership you used to know only as email signatures. Personnel paperwork you didn't know existed until it was due. Training modules that have nothing to do with coordinate systems and everything to do with people. None of it is glamorous. All of it is the actual work. A supervisor who treats the administrative layer as a distraction from the "real" work has it exactly backwards — for your team, you are the administrative layer. You're the one who pushes the reclassification through, clears the request that's been stuck for a month, and translates technical reality upward to the people who make budget decisions. If you don't do it, it doesn't get done, and your team pays for it.

The third thing: you become the context-carrier. As an analyst, you could afford to know your own slice deeply and the rest vaguely. As a supervisor, you're suddenly the connective tissue between programs that don't talk to each other, and the institutional memory for decisions made before half your team arrived. Nobody trains you on this. Start practicing before the title changes: learn what the adjacent teams actually do — not what the org chart says they do.

And the honest one: I almost didn't apply. The voice that says "not yet, one more year, one more certification" is persuasive, reasonable-sounding, and lying to you. Nobody feels ready, because readiness for this job can only be acquired by doing it. The technical foundation matters — your team needs to trust that you understand the work — but your foundation was poured years ago. Apply before you feel ready. The feeling arrives after the offer, not before.

For the young professionals reading this: the analyst-to-supervisor jump is the biggest discontinuity in a GIS career. Bigger than school-to-work. Plan for it the way you'd plan a migration — document everything, test your assumptions, and expect the schema to change.

Hands On

The Promotion Prep System That Actually Worked

No code this week. Instead, the unglamorous machinery behind the Main Event — the system I'd hand to anyone targeting a promotion in the next six to eighteen months. None of it is clever. All of it compounds.

Keep a wins log. One running document, one line per entry, dated. Every shipped automation, every workflow you fixed, every time someone outside your team used your work. Don't editorialize — just log it. When a position posts, you'll have months of receipts instead of a panicked memory exercise the night before the application closes. This one habit did more for me than everything else combined.

Convert wins to STAR stories before you need them. Situation, Task, Action, Result. Take your ten best log entries and write each as a four-sentence story. Government and enterprise interviews are behavioral interviews — "tell me about a time you handled a competing priority" — and the candidates who struggle usually aren't short on experience. They're trying to retrieve it live, under pressure. Pre-compile.

Quantify or qualify — never neither. "Automated a quarterly data update" is forgettable. "Automated a quarterly update that used to take two analyst-days and now runs unattended" is a hiring decision. If you genuinely can't put a number on something, name who depended on it instead.

Write your supervisor's job description for your current team. Starting six months out, keep asking: what would my supervisor have to do for this project to succeed? Then quietly do a slice of it. Coordinate the stakeholders. Write the summary leadership actually reads. Mentor the new hire. When the interview asks about supervisory experience you don't technically have, you'll have the next best thing — a record of supervisory behavior.

From the Community

GPN Volunteer Service Awards — Nominations are open through July 1. If someone in your GIS community has been quietly carrying the load, this is the mechanism for saying so out loud.

2026 COGO NSDI Assessment Survey — Open through June 20. The Coalition of Geospatial Organizations' assessment feeds the national report card on spatial data infrastructure. If you work with framework data, your input belongs in it.

GPN GIS Leadership Academy — On theme this issue: if the analyst-to-supervisor jump is on your horizon, GPN runs a dedicated leadership program built for exactly this transition. Worth a look before you need it, not after.

→ thegpn.org

The Roundup

GeoAI & Robotics Summit — June 16–17, Pittsburgh. The inaugural executive summit on geospatially-enabled autonomous systems: GeoAI, digital twins, advanced positioning. Invitation-only and leadership-focused, but the post-event coverage will be worth tracking for where the autonomy conversation heads next.

AGILE 2026 — June 16–19, Tartu, Estonia. The 29th edition of Europe's GIScience research conference, and reliably one of the better venues for reproducible-research practice in our field.

ML4EO 2026 — June 22–24, Exeter, UK. Machine Learning for Earth Observation. If the Earth Foundation Model conversation from past issues interested you, this is where the next round of papers lands.

Esri User Conference — Next month in San Diego. Start watching the agenda releases now. The plenary is where the year's platform direction gets announced, and the good sessions fill fast.

Coordinates

Every analyst eventually faces the same reprojection: the work stops being the work, and the people become the work. Same ground. New reference frame.

Null Island Dispatch is researched and written with Claude, Anthropic's AI — guided by Chris Lyons, GIS professional and spatial developer. The automation is intentional. The transparency is too. Views expressed are Chris's own and are not affiliated with any employer or organization.