This tutorial shows how to scrape OpenWork job reviews data from company profile URLs into CSV with the OpenWork Job Reviews Scraper for UScraper. You will import the template, replace sample URLs, set the export folder, validate a small batch, and troubleshoot common blank-row issues.
Before you start
Prerequisites, scope, and OpenWork policy checks
You need UScraper installed as a local desktop app, a short list of OpenWork company profile URLs, and a folder where the CSV can be written. Start with three to five URLs so you can validate selectors, row counts, Japanese labels, and salary formatting before running a larger batch.
This guide covers visible company profile summary pages, especially the score block on the company TOP page. It does not cover account-only reviews, candidate profiles, paid datasets, login-gated content, CAPTCHA bypassing, or attempts to defeat OpenWork access controls. OpenWork also offers official alternative data products for investment and research use cases through OpenWork Inc.
Before automation, review the current OpenWork terms of use and OpenWork robots.txt. If your use case involves redistribution, model training, public publishing, or investment research, compare this browser workflow with OpenWork's official alternative data route.
Technical access is not permission. Treat employee-review data as sensitive business research, keep runs modest, and stop when a page asks for registration, verification, or a different access path.
Workflow anatomy
What the OpenWork job reviews scraper exports
The bundled JSON uses a compact multi-URL loop: Navigate -> Wait for Page Load -> Wait for Element -> Sleep -> Structured Export -> Loop Continue. Navigate contains the company profile URLs. The Wait for Element block checks for .averageScore. Structured Export reads from body, writes one row, and Loop Continue advances to the next URL.
This template is not scraping every individual review. It captures company-summary metrics that are visible on the profile page, then keeps the source URL beside the numbers for later audit.
| CSV field | What it captures | Validation check |
|---|---|---|
ページのURL | Current OpenWork company profile URL | Open the link and confirm it is the company you intended to process. |
合計得点 | Total evaluation score from the visible score block | Compare against the score shown beside the company profile. |
待遇面の満足度, 社員の士気, 風通しの良さ | Score-category metrics parsed from page text | Expect blanks if OpenWork changes labels or hides the module. |
社員の相互尊重, _20代成長環境, 人材の長期育成 | Additional workplace evaluation categories | Check that Japanese labels still appear in the rendered page. |
法令順守意識, 人事評価の適切感 | Compliance and evaluation-system score categories | Confirm selector patterns after any site layout change. |
回答者の平均年収, 年収範囲 | Average respondent salary and salary range text | Keep units such as 万円 during downstream cleanup. |
There is no bundled CSV sample, so the JSON export is the authoritative sample. Your first run creates the first real CSV, with headers enabled and append mode on.
Runbook
How to scrape OpenWork job reviews to CSV
Import the template
Open OpenWork Job Reviews Scraper, download the JSON, and import it into UScraper.
Replace sample company URLs
In the Navigate block, replace the three sample company.php?m_id= URLs with your approved OpenWork company profile URLs. Keep one profile URL per input.
Confirm waits
Keep the page-load wait and the visible .averageScore check for the first run. Add time only after you see slow rendering, not as a default habit.
Set the CSV destination
In Structured Export, confirm openwork-job-reviews-cloud-only-scraper.csv, headers, append mode, and a project-specific save folder.
Run one company first
Run a single URL, inspect the row, compare every populated score against the browser, then restore the full URL list and run the batch.
After the dry run, sort the CSV by ページのURL. One OpenWork profile should create one row. If rows repeat, clear the previous file or rename the output before rerunning. If all score fields are blank, the page likely missed the score block or the layout has drifted.
Validation
Validate OpenWork job reviews data before analysis
Do not treat the export as finished when the file appears. Review rows from the beginning, middle, and end of the batch. Open the source URL, compare the total score, confirm salary units, and note companies where OpenWork withheld or changed visible fields.
| Symptom | Likely cause | Fix |
|---|---|---|
| Zero rows exported | URL blocked, page failed to load, or score block never appeared | Run one URL with the browser visible and stop if a gate appears. |
| Blank score columns | Label text changed or company page lacks that metric | Inspect the rendered page and update the JavaScript column pattern. |
| Salary range looks malformed | Japanese range separator or comma handling changed | Normalize only after preserving the original CSV. |
| Same company appears twice | Append mode reused an old file or duplicate URLs were supplied | Deduplicate by source URL before importing into BI tools. |
| Rows look shifted | The CSV was opened with the wrong encoding or delimiter | Reopen as UTF-8 CSV and avoid manual edits before QA. |
Alternatives
Octoparse OpenWork alternative, hosted scrapers, and code routes
If you search for openwork scraping tools or octoparse openwork alternative, you will see several approaches. Octoparse lists OpenWork templates, including a cloud-only job reviews scraper. Spider lists an OpenWork scraper as a hosted extraction option. Developers may also build Selenium or browser-automation scripts for custom parsing or internal pipelines.
UScraper fits supervised CSV export: visible browser QA, editable workflow blocks, and a local file path an analyst can inspect. Hosted scrapers are stronger for scheduling, managed infrastructure, API delivery, and scale. A custom script is stronger for version-controlled parsing logic and test fixtures. Official OpenWork data routes are cleaner when rights, redistribution, and production use matter more than a one-off research spreadsheet.
FAQ
OpenWork scraper tutorial FAQ
OpenWork pages may be publicly visible, but automated collection can still be limited by OpenWork terms, robots directives, copyright, privacy law, and how employee-review data is reused. Review the current rules, keep batches modest, and do not bypass login, registration, CAPTCHA, or other access controls.
Next step
Download the OpenWork job reviews scraper template
When your URL list is ready, download the JSON from OpenWork Job Reviews Scraper and keep this tutorial open for QA. Import the workflow, run one profile, validate the CSV, then scale only as far as your permissions, policy review, and data-quality checks allow.