Why start from postings, then check against O*NET
Course pages describe what a course teaches; job postings describe what employers ask for. A gap analysis compares the second with what you can already show. Five postings for one job title is enough to see a pattern without turning the exercise into a research project.
Postings are noisy, so cross-check them against O*NET OnLine, the occupational database sponsored by the U.S. Department of Labor's Employment and Training Administration. O*NET says it covers over 900 occupations, is free to use and is updated quarterly. Each profile lists tasks, skills, knowledge areas and technology skills.
For example, O*NET's Data Scientists profile (15-2051.00) lists tasks such as analysing large data sets with statistical software, creating visualisations of results, cleaning raw data and presenting results to management or other end users. Its Business Intelligence Analysts profile (15-2051.01) names Microsoft Power BI, Microsoft Excel and several SQL tools, including Transact-SQL and PostgreSQL, among its technologies. If your postings all ask for dashboards but your list has no presentation skills, O*NET's task list tells you that is a gap too.
Step by step: from five postings to a ranked gap list
- Choose one job title and one location or remote setting. Mixing "data analyst" with "data engineer" blurs the tally.
- Save five current postings as text, with the date you saved each. Delete company boilerplate and benefits.
- List every skill, tool and credential named in the requirements. Merge synonyms, such as "dashboards" and "Power BI or Tableau".
- Count how many of the five postings name each item. That count is the demand score, from 1 to 5.
- Rate your current level from 0 to 3: 0 none, 1 followed a tutorial, 2 completed an exercise on my own, 3 built something I can show. Set a target level, usually 2 or 3.
- Gap = target minus current. Priority = demand times gap. Sort by priority.
- Check the top items against the O*NET profile for the occupation and add any core task that every posting assumed but none spelled out.
Worked example: a junior data analyst search
The five postings behind this tally are invented for illustration; replace them with real postings you collect. The arithmetic shows how a ranking emerges.
| Skill | Postings naming it (of 5) | Current level (0-3) | Target | Gap | Priority (demand x gap) |
|---|---|---|---|---|---|
| SQL queries and joins | 5 | 1 | 3 | 2 | 10 |
| Dashboards (Power BI or Tableau) | 3 | 0 | 2 | 2 | 6 |
| Excel: pivot tables and lookups | 4 | 2 | 3 | 1 | 4 |
| Presenting findings to non-technical staff | 4 | 2 | 3 | 1 | 4 |
| Descriptive statistics | 3 | 1 | 2 | 1 | 3 |
| Python for data cleaning | 2 | 1 | 2 | 1 | 2 |
| Bachelor's degree | 3 | Not a study gap | Not applicable | Note separately | Not scored |
Mapping gaps to course modules and projects
Map each top gap to the smallest unit of study that closes it, plus one project that proves the new level. Buying a full program to close one gap is an easy way to overspend.
In the example, SQL is first. Coursera's SQL for Data Science course has four modules; its page estimates 2 weeks at 10 hours a week, while its module estimates add up to about 15 hours. It carries a shareable certificate and is labelled Included with Coursera Plus; Coursera Plus costs $59 a month or $399 a year. Alison's SQL for Beginners is a free, CPD-accredited course listed at 3 to 4 hours, with an optional paid certificate. Alison says its courses are self-paced with no time limit and require at least 80 percent on assessments to pass.
Credentials behave differently. O*NET reports that 68 percent of Business Intelligence Analyst respondents said a bachelor's degree was required. Treat a degree requirement as a separate decision about which roles to target, not as a gap a short course can close.
| Gap | Study unit | Proof project | Time estimate |
|---|---|---|---|
| SQL (priority 10) | Coursera SQL for Data Science (4 modules, about 15 to 20 hours) or Alison SQL for Beginners (3 to 4 hours) | Answer five questions about a public dataset with joins and grouped totals | About 3 weeks at 5 hours a week |
| Dashboards (priority 6) | One dashboard module from a course you already have access to | A one-page dashboard built from the SQL project output | 2 weeks |
| Excel (priority 4) | Pivot tables and lookup lessons only | Rebuild one dashboard figure in a spreadsheet | 1 week |
| Presenting (priority 4) | No course needed | A 5-slide summary of the SQL project for a non-technical reader | A weekend |
Copy-ready template
Paste this into a document or spreadsheet. Keep it next to your study calendar and redo the tally every three months, or when you change target role.
- Target job title / location / date postings saved:
- Posting sources (1-5), with dates:
- O*NET occupation code and profile checked on:
- Skill | Postings naming it (0-5) | Current (0-3) | Evidence for current | Target | Gap | Priority
- Top three gaps and the single study unit for each:
- Proof project for each gap, with finish date:
- Credential requirements noted separately (degree, licence, certification):
Mistakes that distort the result
- Rating yourself on what you have watched rather than what you have built. Write the evidence next to each rating.
- Counting every tool in a long "nice to have" list. Tally required skills first; keep preferred skills in a second column.
- Treating one posting's unusual tool as a trend. An item named in only one of five postings is rarely worth a course yet.
- Studying the easiest gap first because it feels productive. Sort by priority, then start.
- Skipping the proof project. A certificate shows you finished a course; a project shows you can do the task the posting describes.
Questions before you decide
What is O*NET and is it free?
O*NET OnLine is the occupational database sponsored by the U.S. Department of Labor's Employment and Training Administration. It says it covers over 900 occupations, is free to the public and is updated quarterly, with tasks, skills, knowledge and technology skills for each.
How many job postings do I need for a skills gap analysis?
Five postings for one job title is enough to see which skills repeat. Count how many name each skill, rate yourself 0 to 3, and multiply demand by gap. Cross-check the top items with the O*NET profile for that occupation.
Should I buy a whole certificate program to close one gap?
Usually not. Match a gap to the smallest unit that closes it. For SQL, Coursera's SQL for Data Science lists four modules and about 15 to 20 hours, and Alison's free SQL for Beginners lists 3 to 4 hours.
Sources & verification
Product details and prices can change. Check the linked provider before buying.
- O*NET OnLine: Data Scientists (15-2051.00) Accessed 30 Oct 2026
- O*NET OnLine: Business Intelligence Analysts (15-2051.01) Accessed 30 Oct 2026
- O*NET Resource Center: Overview Accessed 30 Oct 2026
- Coursera: SQL for Data Science course page Accessed 30 Oct 2026
- Coursera Plus: plans and pricing Accessed 30 Oct 2026
- Alison: SQL for Beginners course page Accessed 30 Oct 2026
- Alison Help Center: How do Alison courses work? Accessed 30 Oct 2026
Sources link directly to providers. Product buttons may use separately labeled affiliate links. Read our disclosure.