01
Normalize
Convert raw archives into one canonical OHLCV schema with consistent timestamp handling and predictable columns.
Clean Forex DataStop backtesting
Stop backtesting
dirty FX data
HistoricalFX is built around a simple premise: the expensive part of historical forex data is not finding prices. It is cleaning, validating, converting, and documenting them well enough that a trader or developer can trust the result.
What dirty data does to a backtest
Small data defects compound quickly. A single bad spike, shifted session, or duplicate timestamp can change stops, indicators, fills, and performance reports.
Duplicate timestamps create false repeated bars.
Missing minutes break indicators and strategy warmups.
Bad ticks create fake stop-outs or impossible wins.
Broker/session differences shift candles and distort comparisons.
CSV conversion errors silently change dates or numeric types.
Mixed timeframe sources make M1, H1, and daily files disagree.
The HistoricalFX cleaning layer
02
Validate
Check timestamp order, duplicate bars, OHLC relationships, missing files, suspicious outliers, and export readability.
03
Package
Deliver Parquet-first files for Python and modern data tools, with CSV and MetaTrader conversion files rebuilt separately for platform workflows.
04
Document
Publish methodology, release manifests, and known limitations so teams can reason about the data before trusting a backtest.
From cleaned files to data infrastructure
This is the product direction: retail data downloads first, then commercial licenses, release manifests, validation reports, recurring updates, and API access for teams that need market data they can defend.
If your team needs custom forex data cleaning, source comparison, gap detection, or repeatable validation reports, start with a commercial request.
Discuss a data-quality request