Job Market Report — salary percentiles, top skills and remote share from job scraper data

For recruiters, HR and compensation analysts, career coaches, job-board and newsletter owners, and labour-market researchers who already scrape jobs and now have to answer "what does this market actually pay, ask for and allow?" You have thousands of job postings in an Apify dataset but no time to turn them into numbers. This Actor reads the normalized job dataset (the output of Job Postings Normalizer) and writes one market report plus one row per role or country — no scraping, no proxies, no spreadsheets. $0.10 per report + $5 per 1,000 segment rows. Reading and duplicates are free.

What you get

• Salary percentiles p25 / p50 / p75 per currency — annualized (hour × 2,080, day × 260, week × 52, month × 12), using the midpoint of the salary range. Currencies are never mixed or converted.
• Most requested skills — the share of jobs that mention each of about 100 common skills (Python, SQL, AWS, Kubernetes, Salesforce, Figma, Excel …) in the title or description. Add your own tools with extraSkills.
• Remote share — remote / hybrid / onsite counts and remoteShare (remote ÷ jobs with a known workplace type).
• Segments — one row per role family (derived from the job title: Data Analysis, DevOps & Cloud, Sales …), per country, or per role × country, each with its own salary percentiles, remote share, top skills and top hiring companies.
• Report row — jobs analyzed, salary coverage, top roles, top countries, top 10 hiring companies, employment types, posting date range. The same report is saved as one JSON document in the key-value store (REPORT).

Use cases

• Benchmark pay for a role in a country before you write an offer or negotiate
• Show candidates or clients which skills a market asks for and how many roles are remote
• Fill a weekly "jobs market" newsletter or dashboard with fresh numbers from your own scraper runs
• Compare countries or role families in a labour-market study

How to use

1. Run a job scraper (LinkedIn Jobs, Indeed, Glassdoor, career pages …) and note the dataset ID.
2. Run Job Postings Normalizer on it (several datasets can be merged there). It turns salaries and countries into the fields this Actor reads.
3. Run this Actor with the normalizer's dataset ID in datasetIds. Open the first dataset row (type: "report") and the segment rows, or fetch the REPORT record.

Leave datasetIds empty to see a free demo on clearly fictional sample jobs (every row is marked "_demo": true).

Input

{
  "datasetIds": ["<dataset ID of a Job Postings Normalizer run>"],
  "segmentBy": "roleAndCountry",
  "maxSegments": 30,
  "minSalarySample": 3,
  "topSkills": 20,
  "extraSkills": ["dbt", "Terraform"],
  "dedup": true
}

Field	Default	Meaning
segmentBy	role	role, country or roleAndCountry
maxSegments	30 (0–200)	Largest segments first. Also caps segment charges. 0 = report row only
minSalarySample	3	Percentiles are null when fewer salaries of one currency are available
topSkills	20 (1–50)	Skills listed in the report row
extraSkills	none	Extra tools to count (whole-word, case-insensitive)
dedup	true	Same URL, then same company + title + location counted once

Fields read from each job: title, company, countryCode, workplaceType (or remote), salaryMin, salaryMax, salaryCurrency, salaryInterval, postedAt, employmentType, description (only to detect skills; it is not output). A job counts for salary only when it has a currency, an interval and an amount. Datasets that were not normalized still work but give sparse salary and country numbers.

Output

First row: the report.

{
  "type": "report", "jobsAnalyzed": 24, "companies": 7, "countries": 5, "salaryCoverage": 0.833, "salaryUnusable": 1,
  "salaryByCurrency": {
    "USD": {"sample": 9, "p25": 125000, "p50": 147500, "p75": 170000, "interval": "year"},
    "EUR": {"sample": 5, "p25": 75000, "p50": 79000, "p75": 80000, "interval": "year"}
  },
  "workplace": {"remote": 8, "hybrid": 10, "onsite": 6, "unknown": 0, "remoteShare": 0.333, "remoteOrHybridShare": 0.75},
  "topSkills": [{"skill": "SQL", "jobs": 10, "share": 0.417}, {"skill": "Python", "jobs": 7, "share": 0.292}],
  "topRoles": [{"role": "Software Engineering", "jobs": 5}], "topCountries": [{"countryCode": "US", "jobs": 10}],
  "topCompanies": [{"company": "Demo Systems (sample)", "jobs": 5}], "postedFrom": "2026-10-06", "postedTo": "2026-10-14"
}

Then one row per segment (type: "segment"):

{
  "type": "segment", "segmentType": "country", "role": null, "countryCode": "US", "jobs": 10, "jobsWithSalary": 9,
  "salaryCurrency": "USD", "salarySample": 9, "salaryP25": 125000, "salaryP50": 147500, "salaryP75": 170000,
  "salaryInterval": "year", "remoteShare": 0.4, "remote": 4, "hybrid": 4, "onsite": 2,
  "salaryByCurrency": {"USD": {"sample": 9, "p25": 125000, "p50": 147500, "p75": 170000, "interval": "year"}},
  "topSkills": [{"skill": "SQL", "jobs": 4}], "topCompanies": [{"company": "Demo Systems (sample)", "jobs": 2}]
}

In a segment row salaryCurrency and the salaryP* columns show the currency with the most salaries in that segment; all currencies are in salaryByCurrency. Run statistics are in the key-value store as STATS.

Pricing

Pay per event: $0.10 per report (one per run; not charged when no jobs could be analyzed) and $5 per 1,000 segments ($0.005 per segment row). A typical run — one report and 30 segments — costs about $0.25, whatever the number of jobs analyzed. maxSegments and the maximum cost per run in Apify cap the spend; the Actor stops cleanly when the limit is reached. The demo is free.

Notes and limits

• This Actor only reads datasets you already own; it does not access any job site. You are responsible for how the underlying data was collected and used.
• The numbers describe your dataset, not the whole market. A scrape of one job board, one search term or one week is a sample; small segments are noisy, so percentiles need minSalarySample salaries.
• Role families come from keyword rules on the job title (about 25 groups); unusual titles fall into "Other". Skills are keyword matches, so "Go" is counted only as "Golang" or "Go developer/programming" and a skill mentioned in a "nice to have" line counts too.
• Salaries without a currency or interval are left out of percentiles (and counted in salaryUnusable). Annualization uses fixed factors; no currency conversion and no cost-of-living adjustment.
• No personal data is output: no job texts, URLs, applicant or recruiter names. Company names appear only as hiring counts.

More tools from the same developer

• Job Postings Normalizer — unify fields, salaries and dates from any job scraper and remove duplicates
• Job Listings API — every open job from Greenhouse, Lever, Ashby, Workable, SmartRecruiters and Recruitee career sites
• Dataset Merge & Dedup — merge Apify datasets and remove duplicates by the fields you choose
• Google Sheets Export — export an Apify dataset to Google Sheets