One more way to explore
Take a shot at my portfolio.
Play with a copy of this page in Sprite Fusion.
A game by Hugo Duprez / Sprite Fusion ↗
Destroy my portfolio
Mouse or touch to aim.
Phone: landscape. Esc pauses.
Tim Gilbert / Tucson, Arizona
Apps for ecologists, sports fans & nerds.
I’m an ecologist and data scientist who builds R Shiny apps, databases, scrapers, offline field tools, and interactive browser experiences. I like making complicated things easier to use, and a lot more fun. Projects built with assistance from Claude (Anthropic) and Codex (OpenAI).
Ecologist+Data scientist+Builder of odd little worlds
Field identification / TG-14
36.10° N / 112.11° W
FIELD RECORD ACTIVECANIS MODE ACTIVE
SUBJECTTim Gilbert
RANGEMostly Arizona
R SHINY✶PYTHON✶SQL✶CANVAS✶FIELD ECOLOGY✶SPORTS DATA✶OFFLINE-FIRST✶WEIRD SIDE PROJECTS✶
✶✶✶✶✶✶✶✶
Interactive web
Interactive web field experiences.
Explore ecological records and a paired storm model. Keep going for a Tucson Saturday, a mountain road, and a dead spaceship.
Living X-Ray
Peel back a Santa Rita research plot.
Native NEON terrain, aerial and canopy layers, and 179 tagged plant records. Select a plant to inspect its source measurements and model defaults.
Measured terrain · modeled plants · source receiptsExplore the plot ↗
Two Futures
What changes when you dig one basin?
Run two copies of the actual Storm Table model on one clock. Inspect the first sustained difference and save the paired view, ledgers, and recipe.
Time-compressed, uncalibrated exhibit. Not a site forecast.Compare the storm ↗
Scroll + soundThe Long Saturday
One desert day, from first light to kickoff.
Film scrubbing · field-recorded soundscape · four scroll movementsEnter the story ↗
Procedural Canvas27 Miles to Canada
Drive from saguaros to snow in twenty-seven miles.
Six biotic communities · elevation model · live climate shiftStart the climb ↗
Procedural worldThe Orrery
Eleven worlds keep time around one dead ship.
Canvas-baked planets · generative score · shuttle flight · hidden sectorChoose a world ↗Destroy my portfolio
Apps + systems
Field tools and data products.
Offline collection, roster analysis, capture histories, and plant-family comparisons.
OFFLINE · 12 SAFE
01 / FIELD → DATABASE
EcoPlot Mobile
Offline vegetation data collection, cloud sync, office review, and reports. Built for places where the bars on your phone are mostly decorative.
- PWA
- Azure SQL
- Offline sync
Big 12Arizona
vsBig 12Arizona State
Power-4 Girth Index
Compare how all 67 ACC, Big Ten, Big 12, and SEC programs build football and basketball rosters: body maps, position profiles, where the high-school talent actually comes from, and whether recruiting strength turns into wins. Yes, the name is intentional.
- R Shiny
- Live data
- Power-4 matchups
Trap check · 0 of 12
Demo trap line. Tap a trap to check it.
03 / ECOLOGYSmall Mammal Tracker
National trapping records turned into animal histories, home ranges, diversity profiles, and a Hall of Fame for the most-caught little weirdos.
- R
- Mapping
- Ecological stats
Plants in Movies
Explore Arrakis, Middle-earth and Isla Nublar through real plant records. Compare states, look closer with a specimen lens, and save your finds in a field notebook.
- R Shiny
- USDA data
- Field notebook
A few receipts
Real projects. Real records.
Under the hood
I do the database, the interface, and the field logic.
I’m happiest on the small, specific problems, like the offline edit that collided with someone else’s, or the trap tag that got reused. Nobody downstream should ever have to know it happened.
Build the thing
R Shiny apps, progressive web apps, desktop helpers, and interactive browser work.
R · Python · JavaScript · ElectronWire the data
Relational models, offline sync, ETL pipelines, scrapers, and QA/QC that survive real use.
SQL · Azure · SQLite · APIsKnow the field
I spent years on field crews running vegetation protocols. I know what they ask of a crew, and what it is like to enter it all on a tablet.
Vegetation · Wildlife · Spatial dataThe fun part
Maps, charts, motion, and sound. A chart you can poke at (tap a bar, see which families drive the score) beats one you just look at.
Canvas · Plotly · ggplot2 · Audio
Workspace 01 · Santa Rita
Tortoise
24 of 24 tracked animals are on this round’s list.
Workspace 02 · Santa Rita
Invasive plants
246 of 520 patches are on the candidate list.
Workspace 03 · Santa Rita
Restoration
10 of 20 planting sites have a reason for a visit.
All four, running at once
One hub. Three programs. Nothing real to break.
A Shiny demo hub with tortoise, invasive plant, and restoration workspaces, built on generated records so you can poke at everything. Each one gets the same six screens: start here, action center, register, map, outcomes, and data & methods. The desk on the left is a working miniature: pick a workspace and clear its queue.
Before the data
I was collecting data for NEON before I wrote any code to handle it.
Small-mammal trapping, vegetation surveys, electrofishing, and soil work. That’s why I build for one thumb in bad lighting, with no signal.
Send me the weird one
Got a cursed spreadsheet, a strange dataset, or an app idea nobody else understands?
That is probably more interesting than a normal contact form.