QiyanAgent 企小喵实证研究智能体 Download Client
Data Collection Agent

Open a project first, then let the Data Collection Agent handle sources.

The Data Collection Agent organizes sources, calls public APIs, saves raw files, runs sample checks, and hands stable results to cleaning or database work. Use public or authorized interfaces and do not bypass robots, login, access controls, or site limits.

Applies to QiyanAgent V3.5. Last updated: September 14, 2026.

Data Collection Agent steps: confirm that the project is open, choose a collection capability package, review built-in Skills, enter a collection prompt, and send the task.

This is a QiyanAgent V3.5 interface screenshot. The callouts mark capability selection, Skills configuration, task entry, and task submission.

When to Use It

  • Build source lists, collect public datasets, call web APIs, run batch downloads, and archive raw files.
  • Convert collection results into structured CSV, XLSX, JSON, or Parquet files.
  • Sample-check fields, encoding, missing values, and hand usable outputs to DuckDB import.

Steps

  1. Confirm the project is open and decide where collection outputs should be saved in the project folder.
  2. In the Data Collection workspace, review packages for source lists, reproducibility materials, questionnaire design, public data, raw files, AKShare, API collection, site crawling, batch downloads, format conversion, sample checks, and import validation.
  3. Activate a data collection capability manually, or describe the target and let the Agent decide what to load.
  4. To review capabilities, open Settings → Skills and switch to the Data Collection Agent. Built-in Skills sync automatically; use Import only to add collection methods.
  5. Enter a collection prompt with source URLs, target fields, time range, output format, and acceptance criteria.
  6. Click the send button and let the Data Collection Agent organize sources, collect data, save raw files, and write collection notes.

Acceptance Checks

The output should include the source list, raw file locations, structured result files, field descriptions, sample-check results, and cleaning or import suggestions. If a database import is needed, it should name the target DuckDB table and field boundaries.