StockTool.AI 印度股票研究平台：投資組合研究記錄的資料結構建議

A small data contract for a portfolio research log

A portfolio research interface has two dates to preserve: when the underlying information was published and when the user reviewed it. Treating those timestamps as interchangeable makes an old filing look like fresh evidence.

A minimal research record can contain these fields:

• company_id: a stable company identifier
• reporting_period: the period covered by the source
• source_url: a link to the original filing
• source_published_at: the publication date
• reviewed_at: the review date
• claim: the statement that needs verification
• interpretation: the reviewer assessment
• open_question: the uncertainty still to resolve
• status: needs_verification or reviewed

This is a proposed data structure, not a description of any platform implementation.

Keep evidence and interpretation separate

Store the quoted claim separately from the reviewer interpretation. An AI-generated paragraph should not silently become the source of record. Preserve enough location information for a reader to find the statement in the filing.

Append revisions instead of replacing history

When a new filing arrives, create a new review entry. Keep the earlier interpretation and record the reason for the change. This preserves what information was available at each review.

Make missing inputs explicit

A blank source date should display as missing, rather than inheriting the review date. An unavailable filing should leave a claim pending verification.

Test evidence navigation

Ask a reviewer to open a claim, find the source and identify the reporting period. If these steps are difficult, improve the evidence navigation before adding another summary widget.

For readers exploring research tools for India, visit StockTool.AI. This article accompanies that project; the design above is a general suggestion and does not assert that a particular feature is available.

Published for StockTool.AI. Drafted with AI assistance.