Indonesia IT vs Global Remote: Data Notes

Dataset Overview

The Indonesian dataset and global remote dataset were collected independently, then compared across role category, seniority, and skill frequency. Findings 01–04 draw from the 199-listing enriched Indonesian subset. Findings 05–06 draw from skill data across both full datasets.

On the n=2,517 figure in the AI section: The AI gap analysis (28× headline) uses a broader global pool of 2,517 unique listings — the full database before the tech-role filter was applied to produce the 1,010 figure used in Findings 01–06. Title-level AI signal scanning ran across all 2,517 rows; the 8.6% (216 of 2,517) represents AI-titled listings across all role categories, not just tech. The Indonesian figure (0.3%, 3 of 1,039) uses the full Indonesian collection on the same title-level basis. Both sides of the ratio use the same method — title-keyword scan — so the 28× comparison is apples-to-apples within that method.

Note on Indonesian data: The original May 2026 scrape raw data was not archived after the analysis was completed — a process gap we have since corrected. The replicated June 2026 dataset (478 listings, same two sources) is available to download. The global remote full pool (7,001 rows) is exported directly from our live database and covers the same sources and time window used in the research.

Indonesian IT Dataset

Sources

Listings were collected from two Indonesian job boards active in May 2026:

• JobStreet Indonesia — largest general-purpose job board in Southeast Asia, operated by SEEK. Strong coverage of enterprise, corporate, and mid-market Indonesian employers.
• Loker.id — Indonesia-focused job aggregator with strong coverage of SME and regional employers not well-represented on JobStreet.

Both boards were filtered to IT-category listings only. No manual curation was applied to select which listings to include — all IT-categorized listings available during the collection window were included.

What "enriched" means

A listing in its raw form from a job board typically contains: title, company, location, salary range (if disclosed), and a short snippet or category tag. This is enough to count volume and extract titles, but not enough to classify role type, seniority, or skill requirements accurately.

"Enrichment" means fetching the full job description — the complete requirements and responsibilities text as posted by the employer. 199 of the 1,039 Indonesian listings were enriched. This subset was selected to be representative of the full dataset by title distribution before enrichment was applied.

Findings 01–04 (role categories, seniority distribution, and classification gaps) use only the 199 enriched listings, because reliable classification requires the full description. The 840 un-enriched listings are used only for volume and title-level counts where noted.

Global Remote Dataset

Sources

The global remote dataset was sourced from Loker Dollar's own aggregated database of verified global job sources, which in May 2026 included:

• Contra — independent/freelance and full-time remote platform
• We Work Remotely (WWR) — one of the largest remote-first job boards
• RemoteOK — aggregated remote tech listings
• Remotive — curated remote tech listings
• Hacker News Who's Hiring — monthly HN thread, tech-heavy, significant signal for developer and infrastructure roles
• Adzuna — global job aggregator with strong English-language coverage
• The Muse — US-centric platform with company culture pages and remote filtering
• Additional aggregated sources — Loker Dollar indexes other global sources not listed individually here; the 1,010 figure covers all active remote listings across the full indexed set as of May 2026

All 1,010 global remote listings had full job descriptions available at collection time — no separate enrichment step was required for this dataset.

Classification Methodology

Role categories

Each enriched listing was classified into one of six role categories based on the responsibilities and requirements in the job description:

• Software Development — primary function is writing, reviewing, or architecting code
• Management — primary function is leading teams, managing roadmaps, or overseeing delivery
• IT Support — primary function is end-user support, helpdesk, or IT operations
• Infrastructure / DevOps — primary function is systems, networks, cloud, or deployment pipelines
• Data / ML / AI — primary function is data engineering, analysis, or model development
• Other — roles that don't cleanly fit the above: cross-functional IT roles, ERP consultants, IT governance, ICT procurement, digital transformation officers

Classification was applied to each description once. Borderline cases were assigned to the category that best matched the primary function described in the responsibilities section, not the job title.

Seniority levels

Seniority was inferred from the job description text, not the job title. Four levels were used:

• Junior / entry-level — 0–2 years experience required, or explicit "fresh graduate" language
• Mid-level — 3–5 years experience, or equivalent seniority implied by the scope of responsibilities
• Senior — 5+ years experience, or senior-level scope without management responsibilities
• Lead / manager — explicit team management, organizational authority, or budget ownership

Listings that didn't specify experience requirements or scope clearly enough to assign a level were excluded from the seniority distribution figures.

Skills

Skill frequency was counted by scanning job description text for explicit mentions of technologies, languages, frameworks, and platforms. Skills were counted if they appeared as requirements or "preferred" qualifications — not in context like "familiarity with competitors." Counts represent the number of listings that mention the skill at least once; a listing that mentions Python three times is counted once.

Percentage figures in Finding 05 (e.g., "Java: 18.4% of jobs") mean that 18.4% of the relevant dataset's listings mentioned that skill at least once.

Limitations

• The downloadable Indonesian CSV will not reproduce article figures by subcategory mapping. Article role percentages (8% software dev, 33.7% other, 12.1% IT support) were derived from classifying full job descriptions of 199 enriched listings. The downloadable re-scrape CSV contains only job titles and job board subcategory labels — not full descriptions. Applying the subcategory label directly as a role category produces different percentages: a listing tagged "Developer/Programmer" by JobStreet often describes a cross-functional IT coordinator role when you read the full responsibilities text. This discrepancy is not a data error — it is the finding. Indonesian IT job boards systematically mislabel or obscure the primary function of a role. If you run your own analysis on the CSV, expect your software dev % to be higher and your "other" % to be lower than the article figures until full descriptions are fetched and classified.
• Enriched subset is 19% of the Indonesian dataset. The 199 enriched listings are the basis for role, seniority, and "other" category findings. The remaining 840 listings were not classified at this depth. The enriched subset was selected to be representative by title distribution, but it is not a random sample with formal statistical guarantees.
• Global dataset skews English-language and remote-first. Sources like WWR, Remotive, and HN Who's Hiring primarily index companies that self-identify as remote-first and post in English. Companies hiring globally but not through these channels are not represented.
• Job board coverage is not complete. Both Indonesian boards and global aggregators miss listings posted exclusively on company career pages, LinkedIn, or private networks. Volume counts reflect what was indexed, not total market volume.
• Snapshot in time. All data was collected in May 2026. Job market conditions, role distributions, and skill demand shift over time. This dataset reflects one month's active listings.
• Global dataset is 7,001 rows; article used 1,010 tech-filtered. The full dataset (available to subscribers) is the unfiltered pool from 12 indexed sources. The article's 1,010 figure was a tech-role subset (software development, data/ML/AI, security, infrastructure, management). Skill percentages in the full dataset will be lower than article figures because non-tech roles (marketing, operations, content) dilute the denominator. The free 50-row preview reflects the tech-heavy subset. Filter by skills or title keywords to isolate the tech subset in the full download.
• Salary data was not used for findings. Salary disclosure rates differed significantly between the two datasets (Indonesian boards have higher disclosure rates than global remote boards in general). No salary comparisons appear in the article to avoid comparing incompatible disclosure populations.

What we didn't publish

Several data cuts were computed but not included in the article because the signal was too weak or the sample too small to be reliable:

• Industry-by-industry breakdown (banking, fintech, telecom) — sample sizes per-industry in the enriched set were too small for stable percentages
• Regional breakdown within Indonesia (Jakarta vs outside Jakarta) — location data was inconsistently formatted across the two source boards
• Company-size segmentation — not consistently available from job board listings

Questions about this data

If you're citing this research or have questions about specific figures, reach out at contact@lokerdollar.com. We're happy to clarify methodology or point to the specific data behind a claim.

← Back to the article