iOS · App Store
Native Swift app: today’s free forecast, saved hunt pins, maps, wind check, and track. Background location only when the hunter opts in for a sit. Store listing: DeerCast — Prep. Predict. Pursue.
App Store
Fallstudie · Mobile only · Outdoors
DeerCast
DeerCast
Prep. Predict. Pursue
A native hunting app — not a website with a thin wrapper. Hunters open iOS or Android in the dark, see today’s free whitetail forecast for their pin, then decide whether 06:40 or 16:10 is the hour worth sitting.
Was ausgeliefert wurde: native iOS (App Store) and native Android (Play Store). Same Node.js API. PostgreSQL for accounts, pins, maps, and receipts. MongoDB for the hour-by-hour forecast documents and weather ingest. No Super Admin web console on this product — the phone is the product.
iOS
Android
PostgreSQL · MongoDB
07:10 Prime
NW 8 mph
Offline cache
Cedar Ridge, Iowa · CR-WEST pin
Timber · food plot · stand — the area the forecast is scored for
iOS
Native · App Store
Android
Native · Play Store
PostgreSQL
Users, pins, maps, IAP
MongoDB
Hourly forecast documents
Was wir ausgeliefert haben
Mobile only. iOS and Android.
Both stores get the same Node.js API. There is no staff web app on this engagement. If it does not work with gloves, one bar of LTE, and a 5 a.m. battery, it does not ship.
Android · Play Store
Native Kotlin app with the same forecast, map objects, and track flow. Offline map edits write locally, then sync when the ridge has signal again — same contract as iOS.
Play Store
Das Problem in the field
A weather app is not a hunt plan.
Before this build, a hunter with two vacation days stacked a national weather icon, a moon calendar screenshot, and a mapping app that did not know the rut window. “Cold and cloudy” is not an hour. Sitting 05:30–11:00 because the icon looked good burned the only weekday they had.
- No hour resolutionGeneric forecasts said “good morning.” They did not say 07:10–08:40 on this ridge after a 0.4" overnight rain.
- Maps lived in another appWaypoints, wind, and property lines were not sitting next to the movement score. Switching apps in a stand is how phones get dropped.
- No recovery pathA hit at last light meant a forum thread, not a structured track on the same phone that called the sit.
One hunter, one Tuesday
Cole H. · Cedar Ridge, Iowa · pin CR-WEST
- 21:12 Mon — Opens the iPhone in the kitchen. Free today + tomorrow for CR-WEST. Tuesday 07:00–08:30 tagged Prime after overnight rain and a falling barometer. He sets a sit alarm. Nothing to print.
- 05:48 Tue — Walks in on Android (work phone). Forecast still in the on-device cache — ridge has no LTE. Wind check: NW 8. He keeps the west-side pin; east-side food plot would be a swirl.
- 07:22 — Movement window he sat for. Two does, then a mature buck on the field edge. The hour matched what MongoDB scored the night before, not a generic “cloudy and cool.”
- 16:40 — Optional evening sit is Slow. He does not burn the second vacation day. Track is unused. That is still a win.
Forecast
Hour-by-hour movement — for that pin
The engine scores each hour from local temperature, wind, barometric trend, cloud cover, precip, lunar feeding windows, and season (pre-rut / rut / late). Today’s forecast is free. Longer horizons sit behind IAP, same as the stores.
Today · Cedar Ridge
Whitetail forecast
CR-WEST · free day · cached
05:00
Slow
06:00
Build
07:00
Prime
08:00
Prime
09:00
Good
10:00
Slow
NW 8 mph · 34°F · baro falling · moon 28%
Thema für Thema
What the hunter actually does in the app
01 · Predict
Hour-by-hour forecast
Vorher
A 12-hour weather blob. Sit all morning “just in case.”
Nachher
Each hour has a score for that pin. Cole sits 07:00–08:30 and leaves.
02 · Weather
Weather at the hunt pin
Vorher
Phone weather showed Des Moines or the county seat — not the farm 40 minutes out.
Nachher
Temp, wind, barometer, rain, and moon are scored for the saved pin (CR-WEST), not the nearest town.
03 · Map
Waypoints, parcels, wind, plots
Vorher
Hunt plan in one app, map in another, wind in a third.
Nachher
Pins, food plots, wind check, radar, and crop history sit in the same native map.
04 · Track
Hit location → recovery
Vorher
Guess the blood trail from a group text.
Nachher
Mark the hit on an anatomical model; wait/search guidance is on the same phone.
05 · Offline
Ridge with no bars
Vorher
Spinner. Then a screenshot from last night.
Nachher
Last forecast + map objects cache on device. Sync when LTE returns.
06 · Push
Window alerts
Vorher
They remembered to check at lunch. Window was 07:00.
Nachher
Optional push when tomorrow’s Prime hour lands on a saved pin.
Map + track
Scout on the tailgate. Recover after the shot.
Native map layers: property info, hunting land, custom waypoints, wind check, food plots, live Doppler, virtual rain gauges, aerial, and crop history. Offline edits queue locally so a hunter can drop a pin with no service.
Track is a separate flow: tap the hit on the model, log time and weather already on the pin, then follow wait-vs-search guidance. Session documents land in MongoDB; the hunter identity stays in PostgreSQL.
Data stores
PostgreSQL for the hunter. MongoDB for the hours.
Two stores on purpose. Relational data that must not drift (who paid, which pin, which parcel) lives in Postgres. High-write, nested, time-bucketed weather and scores live in MongoDB so an hour document can be replaced without a 24-column update.
PostgreSQL
Source of truth for identity and hunt assets. Node.js talks to it over a connection pool. iOS/Android never open the database — only the API.
- users / devicesApple and Google account links, push tokens, last app version.
- subscriptions / receiptsApp Store and Play IAP, Elite / Pro / Unlimited entitlements, trial windows.
- hunt_locationsNamed pins (CR-WEST), lat/lng, timezone — the key the forecast job joins on.
- waypoints, food_plots, parcel_savesStructured map objects with hunter_id foreign keys.
MongoDB
Document store for anything that arrives as a nested hour array or a radar blob. Indexed on locationId + date.
- forecast_daysOne document per pin per civil day: 24 hour scores, wind, temp, baro, moon, rut phase.
- weather_ingestRaw vendor/NWS observations the scorer consumed — kept for audit when a hunter says “this hour was wrong.”
- radar_frames / rain_gaugesShort-lived tiles the map overlay reads.
- track_sessions / community_postsHit diagrams, photos, feed items — schema varies; Postgres only stores the owner id.
MongoDB · forecast_days · CR-WEST · 2026-11-18
Scored
{
"locationId": "loc_cr_west", // FK in PostgreSQL hunt_locations
"date": "2026-11-18",
"hours": [
{ "h": 7, "score": 92, "label": "Prime", "wind": "NW 8", "tempF": 34, "baro": "falling" },
{ "h": 8, "score": 86, "label": "Prime", "wind": "NW 8", "tempF": 35, "baro": "falling" }
],
"moonPct": 28,
"season": "rut",
"precipInOvernight": 0.41,
"engine": "v3.4"
}
What happens overnight
From weather ingest to a Prime hour on the phone
01
Ingest
Node.js workers pull local weather and moon for every active pin in PostgreSQL. Raw frames land in MongoDB.
02
Score
The movement engine writes a 24-hour document. Prime / Good / Slow is stored per hour, not as a daily slogan.
03
Sync
iOS and Android GET the day for their pins. Today is free. Entitlement for 14-day range is checked in Postgres.
04
Sit
On the ridge the phone reads the cache. Optional push already fired if tomorrow’s Prime hour matched a saved pin.
Technologie
Native phones. Postgres + MongoDB behind Node.js.
This case study is mobile-only. Native iOS and Android talk REST (and a light sync channel) to a Node.js API. PostgreSQL holds hunters, devices, IAP receipts, pins, and map geometry. MongoDB holds forecast documents, weather ingest, radar, track sessions, and the community feed.
Stores are not interchangeable. You do not put 24 nested hour objects in a Postgres row for every pin every night, and you do not put IAP entitlements in a Mongo document that a hunter can fork.
iOS native
Android native
Node.js API
PostgreSQL
MongoDB
Offline cache
iOS & Android
Native forecast, maps, track, push, offline queue. Same API contract on both stores.
Node.js API
Auth, entitlements, forecast read, map CRUD, weather jobs, push fan-out.
PostgreSQL
Users, devices, App Store / Play receipts, hunt pins, waypoints, food plots, parcels.
MongoDB
Hour-by-hour forecast documents, weather ingest, radar, track sessions, feed posts.
Ergebnisse
Was sich nach dem Go-Live geändert hat
Hours, not slogans
Hunters plan around Prime hours on a named pin. Vacation days stop being “sit from dark to 11 because AccuWeather said cloudy.”
One app in the stand
Forecast, wind, parcels, and track share a native shell. No three-app shuffle with gloves on.
Data that can be audited
If a Tuesday score looks wrong, weather_ingest in MongoDB and the pin in PostgreSQL still exist. The model can be retrained; the receipt cannot be guessed.
Bauen Sie Ihres
Need a forecast-class mobile app?
We ship native iOS and Android against a Node.js API, with PostgreSQL for accounts and MongoDB for time-series documents — the same split used on DeerCast.