Market data
Historical prices and volume stored locally, with checks for gaps and quality before anything is analysed.
A personal research workspace for Vietnamese equities: market data, rule-based signals, paper trading, and honest evaluation of what works.
View the source on GitHubI wanted a disciplined way to test investment ideas instead of trusting intuition or a good-looking chart.
The project began as a daily scanner. It grew into a research and operations workspace where I can check data quality, read the market context, compare stock-selection ideas, and record simulated positions.
The goal is a durable, professional research tool. It does not have to produce profit to be useful: a clear answer to "does this idea hold up?" is already a result.
Four parts that run from data to review.
Historical prices and volume stored locally, with checks for gaps and quality before anything is analysed.
Market trend, breadth, relative strength, and rule-based scans that describe the environment before any stock is considered.
Simulated positions with explicit risk rules. No real money, so ideas can be observed honestly over time.
Comparisons of outcomes across dates, budgets, and holding periods, with diagnostics and stated limitations.
A local dashboard to inspect data status, signals, positions, and operational runs.
Daily reports and queries delivered to chat, so the process stays visible without opening the code.
Research is kept apart from the daily workflow.
A hypothesis and the way it will be judged are written down before looking at results.
Rank information coefficients with bootstrap confidence intervals, not a single lucky number.
Exploratory results stay labelled as exploratory. Nothing is presented as a proven strategy.
One layer at a time, avoiding unrelated model families.
quant-stock is a personal research tool. Nothing here is investment advice, a recommendation, or a claim of performance.