Canadian IT Job Market — measured, not estimated

Measured, not estimated

What Canadian tech employers actually ask for

Built by reading real job postings from several sources — including Indeed, LinkedIn, Glassdoor and companies' own applicant tracking systems. No surveys, no articles, no estimates. Every figure on this page comes from counting postings, and every one shows its n so you can judge it yourself.

Ranked by volume

Most in-demand IT roles

The Postings count is the volume each role is measured on. Click any row to see its skills, employers and metros — and how many full-text postings those skills are read from.

Location

	Role	Postings	Remote	Median salary

Share of postings that ask for it

What the market actually wants

Why this number differs from everywhere else

“AI/ML” first measured 70% of all postings. That was false.

The first pass reported AI/ML in 70% of postings — and in 72% of sales roles. Absurd on its face. Inspecting the matches showed what was actually being counted:

425× “…observability and security platform for the AI era…” — one company's About Us paragraph, repeated across all 425 of its postings.
426× “Privacy and AI guidelines…” — a hiring legal notice.

It was measuring marketing copy and legal boilerplate, not job requirements. Text each employer repeats across their own postings was removed (33% of all words), and the bare token “AI” stopped counting on its own. Real AI/ML demand is a fraction of the first figure — still significant, but true.

Where the jobs are

The market by metro area

Grouped by Statistics Canada metropolitan area, because people commute within a metro: a role in Burnaby is a Vancouver job. Counting municipalities separately made Vancouver look smaller than a Toronto suburb.

	Metro area	Postings	Remote	Median	Leading role

What this dataset cannot tell you

Method and limits

A dashboard with honest holes is worth more than a complete one that invents coastline. These are the real limits of what sits behind the numbers.

Sources

Where the data comes from

full text means the entire job description was read, so the posting can feed the skill counts. teaser only sources contribute volume, title, salary and location, but never a skill percentage.

Declared bias

Skills are not measured across the whole dataset

Adzuna supplies most of the postings, but its API returns a 500-character teaser, not the description. You cannot detect “SQL” in a 500-character teaser.

Counting skills across every row deflated every percentage: SQL measured 3.9% dataset-wide (960 of 24,799 rows) against 12.6% on the postings with real text (n=960/7,646). So skills are computed only on postings whose full description we actually have, and that n is always shown.

Duplicates

The same job counted three times

One job is posted to the company's ATS, scraped into an aggregator, and mirrored on LinkedIn: three records, three ids, one real vacancy. Counting all three inflates every figure — and inflates it unevenly, overweighting employers who syndicate hard.

Duplicates are found by matching normalised employer + title + city. were flagged. None are deleted — they stay auditable. One recruiter had posted the same Java role 35 times.

Remote

When “remote” is a fact and when it is a guess

Applicant tracking systems (Ashby, Lever) and several full-text sources carry a real remote flag. The rest do not: there it is inferred from the text, which under-reports.

The average is flatter than the folklore. Remote runs around 19.7% across the market (n=23,142), and product-tech postings sit right at that line — 18.6% (n=6,488). There is no tech premium. The genuinely remote-heavy corners are narrower and elsewhere: Data Analyst at 42% (n=112) and Sales at 29.6% (n=999). Remote work exists; it just isn't concentrated where people assume.

Salary

Partial, and sometimes imputed

Most employers publish no salary. Medians are computed only over postings that disclose one, and that n is always shown — it is usually far smaller than the role's volume.

Adzuna additionally imputes missing salaries. Treat absolute levels with caution; comparisons between roles remain valid, because the same bias applies to all of them.

What cannot be had

Indeed and LinkedIn block direct access

Both return 403 to any automated request, including their RSS feeds. Their postings reach this dataset the legitimate way: through aggregators that already index them.

That means a sample of those two boards, not their full pool. Stated, not hidden.