Erg Master

A rowing tracker application built for rowers, by rowers.

Website Link

Overview

Erg Master is a full stack rowing tracker web application designed to help rowers keep track of their rowing, strength training and flexibility performance, all in one place. Users can record sessions in the app and visualise their progress. A Python-based machine learning model is used to predict future performance. Erg Master can therefore provide rowers with one single place to track all of their progress, rather than having to use separate apps.

Images

Logging

Dashboard

Implemented Features

Logging Rowing Sessions

• Users can log rowing sessions
• Single distance, single time, timed intervals or distance intervals
• Notes can be added also

Logging Strength Sessions

• Users can log strength training sessions
• Users can select exercise from bank (in future will be able to add more)
• Can add sets, reps, weight, rest, RPE

Logging Weight

• Users can log their weight for a specific date, with notes
• A graph of this data can be seen on the dashboard

View History

• Rowing and strength session history can be viewed on history page
• These sessions can also be edited or deleted

Progress Page

• The progress page includes many sections to indictate progress
• Weekly distance
  • user can view their weekly rowing distance, which is tracked on a graph
• Streaks
  • weekly and daily streaks of activity
• Weight Trend
• 2k Performance Predictions

2k Performance Predictions

• An XGBoost model is used to predict performance based on a user's session history -
  1. Total cumulative distance (meters)
  2. Total cumulative training time (seconds)
  3. Number of total workouts logged
  4. Average power output (watts) across all workouts
  5. Max power output (watts) across intervals
  6. Best (fastest) 500m pace in seconds across any interval
  7. Best (fastest) 2000m pace equivalent or 2k interval pace
  8. Recent 14-day training volume (meters)
  9. Recent 14-day average watts
  10. Consistency (standard deviation of daily training frequency or interval count)
• A synthetic dataset is used as training data - this should be replaced by a real dataset in future

Strength Training Templates

• Templates for strength training sessions
• User can set a template with specified exercises and sets
• Can then easily create a workout from a template

Authentication

• User sign-in, sign-up and authentication using BetterAuth
• Including Google sign in

Stack

Frontend

The frontend uses Next.js App Router

• Next.js
• React
• TypeScript
• TailwindCSS

Database

• Neon Serverless Postgres

Backend

Currently, this consists of the XGBoost 2k performance predictor in Python This is accessible through the /predictions endpoint, using FastAPI

• Python
• FastAPI

Development

• Git
• GitHub
• GitHub Actions

Hosting

• Vercel (Frontend)
• Render (Backend)
• Docker

Database Schema

rowing_sessions

Stores one rowing workout for a user. This can be single distance, single time, timed intervals or distance intervals

Field	Explanation
id	Unique identifier for the rowing session.
user_id	Better Auth user who logged the session.
session_date	Date the workout took place.
session_type	Workout format: single distance, single time, timed intervals, or distance intervals.
notes	Optional notes about the workout.
created_at	When the session record was created.

rowing_intervals

Stores the individual pieces of work that make up a rowing session. Each interval belongs to a rowing session. A single distance/time rowing session is modelled as 1 interval Interval sessions are modelled as multiple intervals Split /500m is not stored and is instead calculated

Field	Explanation
id	Unique identifier for the interval.
rowing_session_id	The rowing session this interval belongs to.
interval_number	Interval's position in the session.
distance	Distance rowed, in meters.
time_seconds	Time spent rowing, in seconds.
avg_stroke_rate	Average strokes per minute, when provided.
avg_watts	Average power, in watts, when provided.
rest_time_seconds	Rest after the interval, in seconds; zero for single-piece sessions.
created_at	When the interval record was created.

strength_sessions

Stores one strength workout for a user. Strength sessions are made up of exercises, which in turn are made up of sets

Field	Explanation
id	Unique identifier for the strength session.
user_id	Better Auth user who logged the session.
session_date	Date the workout took place.
notes	Optional notes about the workout.
created_at	When the session record was created.

strength_session_exercises

Lists the exercises performed in a strength session, in workout order.

Field	Explanation
id	Unique identifier for this exercise entry in the session.
strength_session_id	The strength session this entry belongs to.
exercise_id	Exercise performed, referencing exercises.
exercise_order	Exercise's position in the workout.
created_at	When the exercise entry was created.

strength_sets

Stores set-level details for an exercise entry in a strength session.

Field	Explanation
id	Unique identifier for the set.
strength_session_exercise_id	The exercise entry this set belongs to.
set_number	Set's position for that exercise.
reps	Number of repetitions completed.
weight_kg	Weight used, stored in kilograms.
rest_time_seconds	Rest after the set, in seconds, when provided.
rpe	User-reported rate of perceived exertion (0–10), when provided.
created_at	When the set record was created.

exercises

Exercise catalogue shared by workouts and templates. A nullable user_id distinguishes user-created exercises from shared exercises.

Field	Explanation
id	Unique identifier for the exercise.
name	Exercise name.
muscle_group	Muscle group associated with the exercise.
user_id	Owner of a custom exercise; null for a shared exercise.

strength_templates

Stores a named reusable strength-workout template for a user.

Field	Explanation
id	Unique identifier for the template.
user_id	Better Auth user who owns the template.
name	Template name.
created_at	When the template was created.

strength_template_exercises

Lists the exercises included in a strength template and the planned set count for each.

Field	Explanation
id	Unique identifier for this template exercise entry.
strength_template_id	The template this entry belongs to.
exercise_id	Exercise included, referencing exercises.
exercise_order	Exercise's position in the template.
target_sets	Number of sets planned for the exercise.

weight_entries

Stores a user's body-weight measurement.

Field	Explanation
id	Unique identifier for the measurement.
user_id	Better Auth user who recorded the measurement.
weight_kg	Weight value, stored in kilograms regardless of display preference.
recorded_at	Date the measurement was recorded.
notes	Optional notes about the measurement.
created_at	When the measurement record was created.

user_settings

Stores user-specific display preferences.

Field	Explanation
user_id	Better Auth user these settings belong to.
weight_unit	Preferred weight display unit (kg or lb).

Better Auth tables

Better Auth generates and manages the following authentication tables and fields:

Table	Explanation and fields
user	Stores account profile: id (user identifier), name (display name), email (email address), emailVerified (whether the email is verified), image (profile image URL), createdAt and updatedAt (record timestamps).
session	Stores active login sessions: id (session identifier), expiresAt (expiration time), token (session token), createdAt and updatedAt (record timestamps), ipAddress and userAgent (client metadata), userId (the signed-in user).
account	Stores linked sign-in provider credentials: id (account record identifier), accountId (provider account identifier), providerId (identity provider), userId (account owner), accessToken, refreshToken, and idToken (provider tokens), accessTokenExpiresAt and refreshTokenExpiresAt (token expiration times), scope (granted permissions), password (password credential when applicable), createdAt and updatedAt (record timestamps).
verification	Stores verification challenges: id (record identifier), identifier (email or other item being verified), value (verification token/code), expiresAt (expiration time), createdAt and updatedAt (record timestamps).

Planned Features

• Flexibility tracking
• Improved 2k predictions using rowing dataset, not just individual results
• Allow predictions for more distances
• Mobile app
• Import workouts from Concept2