Use J-Quants Data in KabuForge: Your First Research Run

This workflow is for readers who already have J-Quants daily bars saved locally. The goal is to run one public rule and explain which inputs produced which orders and results. API registration and downloading are separate tasks.

The screen instructions follow research courses built from the main public source candidate. They refer to a different version from the stable package installation screens, so check your version first. Consult Get started for environment preparation and the J-Quants quickstart for the corresponding launch method.

Check that your daily-bar CSV or Parquet contains date, security code, open, high, low, close and volume. J-Quants Date, Code, Open, High, Low, Close and Volume fields and the O/H/L/C/Vo aliases are recognized. Decide whether you are using raw or adjusted prices before research, and do not mix different bases in one series.

The public lesson uses codes 4502, 6758 and 8306 as a fixed operational example. It is not a method for choosing your holdings. See price-data checks for dates and missing values, keep the original file, and prepare a research output location.

In Data Center's local market cache, explicitly select your own file or completed manifest. Review securities, dates, price basis, coverage and input identity before validation. Check the declared policies for merging exact duplicates and handling confirmed full-day trading halts.

If rows conflict or missing values lack an explanation, investigate before proceeding to results. Freeze the validated input, verify it again, and save the resulting manifest and SHA256. The shared preparation tutorial provides the screen-level instructions. Freezing identifies the data used in calculation; it does not certify historical availability.

Create a new research strategy in My Strategies and select Price momentum (downloaded data) with your frozen input. The lesson uses 20 observed sessions of momentum, monthly rebalancing, two holdings, JPY 2 million initial cash and a 0.1% one-way fee. Pass preflight, record the conditions, and then start the calculation.

Changing the dates or settings after viewing results creates a separate trial. Finish one run first and learn to read it without changing the rule. See the public momentum sample for order timing and data availability and lookahead for the timing concepts.

In Backtest Results, inspect the input SHA, dates, holdings and cash, fills, skip reasons and fees. In Run History, inspect launch, result, exit status and logs. Do not treat partial outputs from a failed task as completed research.

In the public course, different operations can produce manifests with different bytes even when they select the same rows. Check the receipt associated with this result rather than copying another run's hash. Continue to reading backtest results and preserve the inputs, settings, report and logs as one research bundle.

Finish with a short note covering data source and acquisition date, software version, price basis, fixed conditions, receipt location and unresolved checks. These course results are research-only with pit_guarantee=false. Being able to calculate history does not establish historical availability or future returns.

This operational record does not verify online requests or successful downloads. Acquisition date differs from historical availability time. See point-in-time data for more detail.

Source: the public GUI research courses. This article teaches research procedures; it does not recommend securities or trades or guarantee future returns. See the disclaimer.

Explore the j-quants learning path

• What Is J-Quants? Research Uses and Limits
• J-Quants Price Data: Fields and Observation Dates
• J-Quants Financial Data: Periods and Units