Swellpoint Co.
Surf Intelligence
Washington Pier
Venice Beach · CA
swell point
A personalized surf intelligence system the beacon that calls the crew to the stoke.
◉ 001 / Architecture
Five
stages,
one
beacon

Post-session data collection flows into a personalized prediction model. Each stage feeds the next — the loop gets smarter with every session logged by the crew.

01

Collect.

Post-surf poll via SMS or email. Rate your session, log your board, note how conditions felt. 2–5 surfers contributing data to the crew model.

02

Ingest.

Pull Surfline data for Washington Pier at session time — swell height, period, direction, wind, tide, temps, crowd, rating. Auto-matched to your poll by timestamp.

03

Correlate.

Match subjective ratings to objective conditions. Build a preference model — per surfer, and an aggregated crew consensus.

04

Predict.

Score upcoming Surfline forecasts against your preference model. Surface the windows where conditions match your best-ever sessions.

05

Notify.

Push alerts when a high-match window lights up. Feed into your weekly planner. Email the Swellpoint crew digest.

Core
components / six

06 Modules
01 — Collect
Session
Logger
Mobile-first poll form sent via Twilio SMS or SendGrid email after each session. Quick 60-second completion. Logs vibe, wave quality, crowd feel, board used, wetsuit worn, and optional notes.
02 — Ingest
Surfline
Pipeline
Scheduled pull from Surfline API for Washington Pier. Captures swell height, period, direction, wind speed/direction, tide, water temp, air temp, weather, and Surfline’s own rating. Historical + forecast.
03 — Engine
Correlation
Engine
Weighted scoring model that learns which conditions map to your best sessions. Tracks personal preferences and crew consensus. Improves with every logged session.
04 — Gear
Board &
Suit Advisor
Maps your quiver to conditions. Tracks which board per session and how it felt. Wetsuit rec based on air + water + wind chill + sun. Over time, tells you exactly what to grab.
05 — Alert
Alert
System
Scans the 7-day forecast against your model. Sends a paddle out ping when conditions score above threshold. Includes board rec, wetsuit rec, time window. Quiet hours configurable.
06 — Plan
Planner +
Team Digest
Weekly forecast piped into Google Calendar. Team email digest with top windows, who’s planning to go, and crew stats. Keeps the Swellpoint crew synced.
33.9758°N · 118.4731°W
Breakwater
Every
session
is a pair.

Your subjective poll meets Surfline’s objective conditions at the same timestamp. That pair is the atom of the system.

Stack enough of them, and a personal model emerges — one that knows your Washington Pier better than Surfline ever could.

FieldSourceTypeNotes
Wave HeightSurflineft (min/max)Primary swell at Washington Pier
Wave PeriodSurflinesecondsLonger period = more power
Swell DirectionSurflinedegrees / cardinalSW vs NW vs S — huge at WP
Wind SpeedSurflinemphOffshore < 5mph = glass
Wind DirectionSurflinedegrees / cardinalNE/E = offshore at Venice
Tide LevelSurflineft + dirMid-tide often best at WP
Water TempSurfline°FFeeds wetsuit recommendation
Air TempSurfline°FComfort factor pre / post session
WeatherSurflineconditionSunny, overcast, fog, etc.
Surfline RatingSurflineFLAT → EPICTheir call — we build a better one
FieldSourceTypeNotes
Session VibePoll1–5Overall, how was it?
Wave QualityPoll1–5Shape, power, rideability
Crowd LevelPoll1–51 = empty, 5 = zoo
Board UsedQuiverselectWhich board you rode
Board FeelPoll1–5Right board for today?
Wetsuit WornPollselect3/2, 4/3, spring, trunks
Comfort LevelPoll3-optionToo cold / just right / warm
Session TimePolldatetimeWhen you paddled out
DurationPollminutesHow long you stayed out
NotesPollfree text“Barrels on the inside”
OutputSourceLogic
Swellpoint ScoreDerivedWeighted correlation of conditions → session vibe
Feels-Like TempDerivedWater + air + wind chill + weather → comfort rating
Wetsuit RecDerivedFeels-like temp mapped to crew comfort reports
Board RecDerivedConditions → board-feel scores from logged sessions
Surfline DeltaDerivedWhere Surfline disagrees with your actual experience

How
was
your sesh?

Sent via SMS or email right after your session. Under 60 seconds on your phone, sand on your hands and all. Every answer feeds the model.

Log Session
Washington Pier · Today 7:15 AM
Overall Vibe
1Kook
2Meh
3Solid
4Fire
5Epic
Wave Quality
1
2
3
4
5
Crowd
1Empty
2
3
4
5Zoo
Board
Shortboard
Fish
Mid
Log
Right Board?
1Nope
2
3
4
5Yes
Comfort
Too Cold
Just Right
Warm
Wetsuit
Trunks
Spring
3/2
4/3
◉ Delivery
How it lands.
Trigger: A manual “just surfed” text, or a scheduled ping at your usual dawn patrol time.

Channel: SMS link via Twilio or email via SendGrid. Tapping opens the mobile web form.

Fallback: If no response in 4 hours, a gentle nudge: “Still dripping? Log your sesh.”

Auto-fill: Date, time, and spot are pre-filled. Surfline conditions auto-attached on the backend.
◉ Crew Mode
Multi-surfer from day one.
Each crew member gets their own poll link tied to their profile. Individual preference models plus a crew consensus aggregate. Board quivers are per-surfer. The team digest shows everyone’s ratings so you can calibrate against your crew.

The
engine
knows you.

Turning every logged session into a personalized forecast — smarter than Surfline’s generic rating, because it’s built on your actual sessions at this actual break.

01
◉ Scoring
Swellpoint
Score
A 0–100 score for any forecast window, personalized to you. Weighted by how much each condition variable correlates with your high-rated sessions.

Example: If your best sessions cluster around SW swell 3–5ft, 12s+ period, light offshore, mid-tide — a matching forecast scores 85+. Cold start uses Surfline rating as baseline for the first 10 sessions.
02
◉ Comfort
Feels-Like
Model
Combines water temp + air temp + wind chill + sun/overcast into a single comfort score.

Calibrated by crew comfort reports — if everyone in a 3/2 says “just right” at 60° water / 65° air / 5mph, that’s the baseline.

Output: “Trunks day” / “Spring suit” / “3/2 full” / “4/3 + boots”
03
◉ Quiver
Board
Recommender
Tracks board × conditions × feel rating. Builds a condition → board fit map for your quiver.

2–3ft, long period, glassy → Fish (4.8 avg)
4–6ft, short period, onshore → Shortboard (4.2 avg)
1–2ft, mushy → Mid-length (4.6 avg)
04
◉ Delta
Surfline
Delta Tracker
Tracks where Surfline’s rating disagrees with your experience. Maybe they call it “Fair” but your crew rates it great — or vice versa.

Builds a correction layer: “Surfline says Fair+ but Swellpoint says this is actually your kind of day.”
10
Sessions to Activate
30
For Solid Model
1yr
Seasonal Patterns
2–5
Surfers for Consensus
Washington / Venice Pier
◉ Where it flows

Signal
to shore.

01

Weekly
Planner

Top surf windows in your calendar.
Format: Calendar events for windows 70+
Includes: Time, score, board, wetsuit
Platform: Google Calendar / iCal
Cadence: Sunday + daily refresh
02

Team
Digest

Weekly email to the crew.
Includes: Best windows, crew stats, deltas, leaderboard
Platform: SendGrid / Gmail API
Cadence: Weekly + real-time alerts
03

SMS
Alerts

Paddle-out pings in real time.
Trigger: Score crosses 75+
Message: “WP fire — SW 4ft @ 14s, glass. Grab the fish. 3/2.”
Platform: Twilio
Controls: Quiet hours, max/day
04

Surfline
Feed

The data firehose.
API: Surfline, WP-specific
Frequency: 3h current, daily forecast
Storage: Postgres time-series
Fallback: Cache + scrape
05

Dashboard

Visual hub for the crew.
Views: 7-day forecast, session log, preference heatmap, board chart
Platform: Next.js + Vercel
Access: Team login, mobile-first
06

Quiver
Profile

Know the boards, know the recs.
Per board: Name, type, length, volume, fins, photo
Purpose: Recs say “grab the 5’8 Pyzel” not just “shortboard”
Setup: One-time, updatable
El Porto
Topanga
Breakwater

Build
in phases.

Phased rollout — start logging sessions immediately, build intelligence over time. Every session makes it smarter.

01
◉ Phase 01
Poll + Data
Build the post-surf poll web form. Set up Twilio SMS. Create the session database. Start logging immediately — even before Surfline integration, you’re building the dataset.
Weeks 1–2
02
◉ Phase 02
Surfline Ingest
Connect Surfline API for Washington Pier. Auto-match conditions to logged sessions by timestamp. Build historical conditions database. Scheduled pulls for forecast data.
Weeks 2–3
03
◉ Phase 03
Correlation Engine
Build the weighted scoring model. Generate Swellpoint Scores for forecasted conditions. Implement feels-like temp model and wetsuit recs. Need ~10 sessions to activate.
Weeks 3–4
04
◉ Phase 04
Alerts + Planner
SMS paddle-out alerts when Swellpoint Score spikes. Google Calendar integration for weekly forecast. Board recs from quiver profile.
Weeks 4–5
05
◉ Phase 05
Team + Dashboard
Weekly email digest to Swellpoint crew. Web dashboard with forecast, session history, model insights. Crew consensus model alongside individual preferences.
Weeks 5–6
06
◉ Phase 06
Refine + Scale
Model improves with every session. Add more spots beyond Washington Pier. Seasonal patterns after one year. Open to other surfers and crews.
Ongoing
Tech / stack
Frontend
  • Next.js
  • Tailwind
  • Vercel
Backend
  • Node / Python
  • Supabase
  • Surfline API
Messaging
  • Twilio
  • SendGrid
  • Gcal API
Intelligence
  • Pandas / sklearn
  • Weighted correlation
  • Cron scheduler
◉ Get the app

Live cam, always with you.

Wave height alerts Right in your pocket.
Watch the lineup Download on the App Store