Case Study
Code Station

TraderBase

Trading journal and analytics

I outgrew tracking trades in spreadsheets and Evernote, so I built one place to log every trade, measure it in risk rather than dollars, and review it honestly.

Role
Design and development
Year
2026
TraderBase monthly dashboard showing current mode, total R, win rate and average risk per trade
Dashboard: the month in R, with the current mode and the distance to promotion or demotion.

01

What it does

TraderBase is my trading journal and analytics app. It logs every equity and futures trade, measures performance in R (multiples of the risk taken) and runs a mode engine that moves me between Defense, Mid and Attack based on recent results. Around the trade log sits the rest of the journal: playbooks, a loss journal, periodic reviews and a research log of chart studies.

  • Everything in R

    Each trade is scored as a multiple of the risk taken, so results compare cleanly across position sizes and account sizes.

  • Mode engine

    Recent results promote or demote me between Defense, Mid and Attack automatically. I don't get to decide how aggressive to be.

  • Journal and reviews

    Playbooks for each setup, a loss journal, and weekly, monthly and annual reviews, all linked back to the trades.

  • Broker reconciliation

    Execution data from Interactive Brokers is pulled in and checked against the journal, so the recorded P&L matches the real one.

02

Screens

TraderBase review mode showing one trade with its chart, result in R and pre-trade note
Review mode: one trade at a time, with its chart, result and the note written before entry.
TraderBase monthly review form with fields for what worked, what didn't, rule violations and goals
Monthly review: what worked, what didn't, rule violations and goals for the next month.
TraderBase historical deep dive showing a chart study with notes on individual stocks
Historical deep dive: a research log of chart studies from past markets.

03

How I built it

TraderBase is a Flask backend with a single-file front end in plain HTML and JavaScript, backed by SQLite. It is hosted on Railway and deploys automatically on every push to GitHub. I built it with AI-assisted coding: Claude for planning and design decisions, and Claude Code to write the changes, one scoped instruction at a time, with every working state committed to git.

Back end
Python, Flask
Data
SQLite
Front end
A single HTML file with plain JavaScript. No framework
Hosted on
Railway, auto-deploying from GitHub
Connects to
Interactive Brokers execution data, and Keel as an optional sync
Built with
Claude and Claude Code