# Data Analysis Work in 2026: Evidence from 126 Current Vacancies

> A reproducible analysis of 126 current public vacancies maps the querying, data quality, reporting, visualization, documentation and communication work employers assign to data analysts.

- Canonical page: https://mtfinstitute.com/insights/data-analysis-work-126-vacancies-2026/
- Content type: Article
- Editorial category: Research & Reports
- Publisher: MTF Institute of Management, Technology and Finance
- Author: MTF Institute Research Team- Published: 2026-08-31
- Updated: 2026-08-31
- Language: English
- Topics: Data Quality, Vacancy Analysis, Data Analysis, SQL, Dashboards and Reporting, Insight Communication

## What 126 Current Public Vacancies Reveal About Practical Data Analysis Work in 2026

The complete open archive - a 10-page visually reviewed PDF, the rights-reviewed 126-row coded dataset, quality records, methods and data dictionary - is preserved at [Zenodo DOI 10.5281/zenodo.22206454](https://doi.org/10.5281/zenodo.22206454). The direct public PDF is [available here](https://zenodo.org/records/22206454/files/data-analysis-work-126-vacancies-2026.pdf?download=1).

**MTF Research Report**  
**Evidence cut-off:** 31 August 2026  
**Corpus:** 126 accepted current public vacancies  
**Course research key:** `professional-certificate-data-analysis`

## Executive summary

This report examines 126 current public vacancies accepted through a two-part, deduplicated evidence process on 31 August 2026. The purpose is not to estimate the size of the labour market or predict hiring outcomes. It is to identify the practical skills, recurring tasks and professional outputs that employers visibly associate with data-analysis work across a deliberately varied set of roles, locations, industries and applicant-tracking systems.

Three findings dominate the corpus. First, analysis is expected to produce usable reporting: 97 vacancies, or 77.0%, carried the dashboard/reporting code. Second, querying remains a central technical capability: 88 vacancies, or 69.8%, explicitly signalled SQL. Third, the work does not end with a correct calculation. Insight communication appeared in 74 vacancies, or 58.7%. Data cleaning and quality appeared in 51 vacancies (40.5%), while ad hoc, diagnostic or deep-dive analysis appeared in 42 (33.3%).

The evidence therefore supports a practical definition of professional data analysis as a traceable chain: clarify a business question, obtain and query relevant data, assess its quality, create suitable measures, perform proportionate analysis, communicate results through reports or visual evidence, document assumptions and limitations, and hand a recommendation to an authorized decision-maker. A course built only around software commands would miss much of the observed role. A course built only around storytelling would be equally incomplete.

The corpus is broad but not statistically representative. It contains 122 unique employers and nine source families. The United States accounts for 54 vacancies (42.9%) and India for 16 (12.7%). No employer contributes more than two vacancies. These properties reduce dependence on any single employer, but the sample remains purposive, English-indexed and weighted toward organizations using public digital recruitment systems.

Responsible use of artificial intelligence (AI) should be bounded. Only two accepted records (1.6%) explicitly mentioned AI-assisted analytical work in the retained observable skill fields. Because the ledger preserves minimal evidence, this is not a prevalence estimate. It does, however, provide no basis for making AI the identity of the professional role. AI can assist with question refinement, draft queries, critique and quality checks, but retained findings, calculations, citations and consequential recommendations require human verification.

## 1. Purpose and research questions

The report addresses five questions:

1. Which technical capabilities appear most often in the accepted vacancy corpus?
2. Which tasks recur across general, business-intelligence, product, marketing, finance and operations contexts?
3. Which outputs make analysis useful to employers and stakeholders?
4. What do these patterns imply for an introductory professional learning pathway?
5. What limits, rights controls and responsible-AI boundaries must accompany that interpretation?

The unit of analysis is one accepted public vacancy, not one mention of a term. A vacancy can receive multiple codes. Percentages therefore describe the share of 126 vacancies carrying a code; they are not intended to add to 100%.

## 2. Evidence and method

### 2.1 Sampling frame

The accepted corpus combines two independently assembled vacancy datasets. Dataset A contributed 60 records from Greenhouse, Lever and Ashby. Dataset B contributed 66 records from Workday, SmartRecruiters, Ashby, UKG Pro Recruiting, Workable, Manatal Careers Page and Oracle Recruiting Cloud. All pages were public and observable without authentication on the evidence cut-off date.

The sampling design was purposive and sought maximum practical variation. Searches covered Data Analyst, Business Intelligence Analyst and closely related reporting, insights, product, marketing, finance, commercial and operations titles where analysis was central. The collection deliberately expanded beyond one applicant-tracking-system family and sought geographic and industry diversity.

### 2.2 Inclusion and exclusion

A vacancy was included when the visible role centered on one or more of the following: querying data, cleaning or validating it, defining metrics, applying analytical or statistical methods, building reports or dashboards, investigating business questions, or communicating decision-ready findings.

The researchers excluded data-engineering-only, data-science-only and software-engineering roles; data-entry and administrative processing positions; future pipeline or proposal roles; vacancies with an expired stated deadline; redirected or closed pages; and records without enough observable evidence to code safely. Hybrid titles were accepted only when practical analysis remained a substantial responsibility.

### 2.3 Current-status evidence

Current status was a point-in-time observation. It was supported by a public Apply, Join or Submit control, a rendered application form, explicit current/open wording, or a recent visible posting date without a passed closing date. This method shows that the vacancy was publicly observable at retrieval. It does not guarantee that the employer would continue accepting applications after 31 August 2026 or that the vacancy had not been informally paused.

### 2.4 Deduplication and acceptance

The merged corpus was checked first for exact URL duplication and then for normalized employer, title and location duplication. The validation file reports 126 inputs, 126 accepted records, no rejected rows, no exact-URL duplicates and no normalized employer-title-location duplicates. All 126 records have a source dataset ID, source vacancy ID, URL, employer or publisher, title, location, source family, retrieval date, current-status evidence, supporting excerpt, observable skills or responsibilities, taxonomy codes, suitability reason and deduplication key.

No employer dominates the evidence. The 126 records represent 122 unique employers. Four employers have two vacancies each, the observed maximum. A notional top-five employer count is nine records, or 7.1% of the corpus.

### 2.5 Coding

Coding was conservative and multi-label. A code was assigned only when a capability or responsibility was explicitly named or clearly observable in the retained vacancy evidence. Dataset A's original skill signals were mapped deterministically to the shared taxonomy; Dataset B's existing taxonomy codes were preserved. Counts indicate vacancies carrying a code, not the number of term occurrences.

The shared taxonomy covers:

- technical capabilities: SQL, spreadsheet analysis, Power BI, Tableau, Python or R, cloud warehouses, extraction/transformation, statistics/forecasting and data modelling;
- professional tasks: data cleaning/quality, reporting/dashboard work, ad hoc analysis, metric definition, requirements discovery, insight communication, documentation/governance, process improvement and experimentation;
- context tags: finance, marketing, operations and product.

### 2.6 Rights and source handling

Only public pages were used. The evidence ledger stores URLs, derived facts and one minimal supporting excerpt per vacancy rather than full copyrighted descriptions. The longest accepted excerpt is 14 words; none reaches 25 words. Dataset A's method classifies its 60 sources as employer-controlled ATS pages. Dataset B explicitly classifies 61 records as first-party employer vacancies and five as public staffing/recruitment vacancies.

The public availability of a vacancy does not grant permission to reproduce its full text, branding, screenshots or application data. This report therefore paraphrases patterns and links to a selected set of source pages for audit. Staffing records are retained as skills evidence but are not used to infer the identity or hiring volume of undisclosed client employers.

## 3. Corpus composition

### 3.1 Source families

| Public source family | Vacancies | Share |
|---|---:|---:|
| Greenhouse | 40 | 31.7% |
| Workday | 24 | 19.0% |
| SmartRecruiters | 22 | 17.5% |
| Ashby | 19 | 15.1% |
| Lever | 10 | 7.9% |
| UKG Pro Recruiting | 5 | 4.0% |
| Manatal Careers Page | 2 | 1.6% |
| Oracle Recruiting Cloud | 2 | 1.6% |
| Workable | 2 | 1.6% |

The source distribution improves platform diversity compared with a single-board sample, but it is not a measure of employer adoption or vacancy volume by ATS provider.

### 3.2 Geography

The two largest normalized geography groups are the United States with 54 vacancies (42.9%) and India with 16 (12.7%). The Philippines and the United Kingdom each contribute five (4.0%). Australia contributes four (3.2%). Canada, Germany and Thailand contribute three each (2.4%). France, Indonesia, Italy, Mexico, Morocco, South Africa and a combined United States/Canada label contribute two each (1.6%). Four records (3.2%) have no normalized jurisdiction, while the remaining records cover Australia/New Zealand, Brazil, Colombia, Egypt, Europe, Finland, Latin America, a multi-jurisdiction remote group, the Netherlands, Nigeria, Singapore, South Korea, Sweden, Ukraine and a United States publisher context.

This spread demonstrates that the identified capability bundle is not confined to one country in the corpus. It does not establish equal demand, comparable seniority or equivalent working conditions across countries.

### 3.3 Employer concentration

The low employer concentration is useful for curriculum interpretation. A recurring code is less likely to be an artifact of one employer repeatedly publishing similar roles. Even so, the sample can still share sector and platform biases because employers using public English-language ATS pages are not a random cross-section of all organizations.

## 4. Findings: technical capabilities

| Technical code | Vacancies | Share of 126 |
|---|---:|---:|
| SQL | 88 | 69.8% |
| Power BI | 41 | 32.5% |
| Tableau | 40 | 31.7% |
| Data modelling | 32 | 25.4% |
| Cloud warehouse/platform | 31 | 24.6% |
| Python or R | 31 | 24.6% |
| Statistics or forecasting | 30 | 23.8% |
| ETL/transformation | 25 | 19.8% |
| Excel or Google Sheets | 25 | 19.8% |

### 4.1 SQL is the clearest common technical language

SQL appears in 88 vacancies, nearly seven in ten records in this corpus. The finding supports practical instruction in selecting data, filtering, aggregation, joins, conditional logic, null handling, common table expressions or subqueries, and validation queries. It does not imply that every analyst works directly against a production database. Safe professional learning should use a small synthetic relational schema and separate query reasoning from production access, security administration or platform engineering.

### 4.2 Visualization tools are important, but the method matters more than one interface

Power BI appears in 41 vacancies and Tableau in 40. Because codes overlap, these counts cannot be added to infer a combined visualization-tool share. The near-equal observed counts support tool-transferable learning: choose an appropriate chart, define a metric, structure a view, label units, avoid misleading scales, annotate findings and test accessibility. A learner can then apply these principles in the employer's chosen tool.

### 4.3 Modern platforms coexist with foundational analysis

Data modelling appears in 32 vacancies, cloud warehouses or platforms in 31, and ETL/transformation in 25. These signals indicate that analysts increasingly work within shared data systems and must understand grain, relationships, semantic definitions, transformation logic and lineage. They do not justify turning an introductory course into data engineering or warehouse administration. The professional boundary is to reason about the data used for analysis, document transformations and identify when a platform or engineering specialist is required.

### 4.4 Programming and statistics are valuable extensions, not the whole role

Python or R appears in 31 vacancies and statistics/forecasting in 30. The evidence supports portable analytical reasoning and introductory scripted analysis, while keeping the learning outcome proportionate. Learners should be able to summarize distributions, compare groups, reason about sampling and uncertainty, interpret associations and simple tests, and check assumptions. The corpus does not support a promise of machine-learning specialization or production model deployment.

### 4.5 Spreadsheet fluency remains relevant

Excel or Google Sheets appears in 25 vacancies. This is lower than SQL in the coded corpus, but it remains a material professional signal. The appropriate response is not vendor-exam preparation. It is transferable spreadsheet reasoning: data types, filters, portable formulas, lookups, error checks, summaries, pivot-style aggregation and a clear calculation record.

## 5. Findings: recurring tasks and professional outputs

| Responsibility or output code | Vacancies | Share of 126 |
|---|---:|---:|
| Dashboard/reporting | 97 | 77.0% |
| Insight communication | 74 | 58.7% |
| Data cleaning/quality | 51 | 40.5% |
| Ad hoc/deep-dive analysis | 42 | 33.3% |
| KPI/metric definition | 36 | 28.6% |
| Stakeholder requirements | 34 | 27.0% |
| Documentation/governance | 33 | 26.2% |
| Experimentation | 18 | 14.3% |
| Process improvement | 17 | 13.5% |

### 5.1 The dominant output is a report, dashboard or visualization that can be used

Dashboard/reporting is the most frequent code in the entire taxonomy: 97 vacancies, or 77.0%. This does not mean that every role is a dashboard developer. The descriptions include recurring reports, executive views, operational dashboards, visualizations and self-service analytical products. The common requirement is a stable evidence surface that helps another person monitor performance, investigate a question or make a decision.

An effective learning pathway should therefore require more than exploratory work. Learners should define the audience and decision, select a limited set of measures, establish refresh and quality expectations, present comparison and uncertainty clearly, and explain what action the output can and cannot support.

### 5.2 Communication is part of the analytical method

Insight communication appears in 74 vacancies (58.7%). Employers seek people who can explain findings to technical and non-technical stakeholders, prepare presentations, recommend action or improve data literacy. Communication is not a decorative final step. It influences which question is asked, which metric is credible, which caveat matters and how the analysis will be used.

A professional output should distinguish observed facts, calculations, interpretations, hypotheses, limitations and recommendations. This distinction reduces the risk that a visually polished result is mistaken for stronger evidence than the data supports.

### 5.3 Trust depends on quality controls

Data cleaning and quality appears in 51 vacancies (40.5%). The observable work includes validation, reconciliation, anomaly investigation, data integrity, duplicate handling and report QA. A beginner should therefore learn to create a data dictionary, quality log and transformation record; inspect missingness, invalid ranges and inconsistent categories; reconcile totals; and preserve a reproducible trail from source to output.

Quality work is not the same as enterprise governance. Analysts need to identify fitness for use and document limitations. Ownership structures, access policy, platform security, formal compliance and organization-wide governance programmes belong to authorized specialists and adjacent learning pathways.

### 5.4 Analysts move between planned reporting and open questions

Ad hoc or deep-dive analysis appears in 42 vacancies (33.3%). KPI and metric work appears in 36 (28.6%), and stakeholder requirements in 34 (27.0%). Together, these codes show that analysts must handle both recurring measurement and less structured investigation. They need to convert a vague request into a bounded question, define the unit of analysis and evidence cut-off, select defensible measures, examine alternative explanations and produce an answer that is timely without overstating certainty.

### 5.5 Documentation is a professional deliverable

Documentation/governance appears in 33 vacancies (26.2%). Observable outputs include data dictionaries, metric definitions, lineage, assumptions, report logic and usage guidance. Documentation makes analysis reviewable and reusable. It also enables a future analyst to reproduce, challenge or update the work without relying on private memory.

### 5.6 Experimentation and process improvement are meaningful but secondary

Experimentation appears in 18 vacancies (14.3%), while process improvement appears in 17 (13.5%). These findings justify introductory coverage of measurement plans, comparison groups, experiment interpretation and analytical workflow improvement. They do not support causal claims without an appropriate design or guaranteed performance improvements.

## 6. Implications for professional learning

### 6.1 Teach one connected analysis-to-decision workflow

The evidence supports a course organized around a realistic business question and one synthetic dataset rather than disconnected software demonstrations. A defensible sequence is:

1. define the decision, stakeholders, measures, data grain and evidence cut-off;
2. inspect the dataset and record data definitions;
3. clean, validate and reconcile the data;
4. use spreadsheet and SQL logic to create reproducible analytical tables;
5. apply descriptive statistics and proportionate uncertainty reasoning;
6. define KPIs and comparisons that match the question;
7. create accessible visual evidence and a decision-oriented report;
8. communicate findings, alternatives, limitations and next steps;
9. preserve documentation and quality checks for review and handoff.

### 6.2 Make the principal output a decision-support package

The recurring vacancy outputs support a capstone such as a Data Analysis Decision Pack. It can combine a question-and-measure brief, data dictionary, quality log, reproducible spreadsheet or SQL analysis record, statistical interpretation, visual evidence, executive narrative, limitations and a next-step recommendation. This is more faithful to the corpus than a collection of unrelated charts or code snippets.

### 6.3 Keep the course tool-agnostic and artifact-led

SQL, Power BI, Tableau, spreadsheets, Python/R and cloud platforms all appear, but no single product represents the whole profession. Learning should therefore teach portable concepts and use compact, original examples. Product names may be referenced factually without logos, trade dress, endorsement or a vendor-certification promise.

### 6.4 Separate analysis from adjacent specialisms

The safe scope is bounded business data analysis. It excludes production database administration, data engineering, machine-learning engineering, regulated financial or clinical conclusions, automated hiring or credit decisions, and claims of causal impact without an appropriate design. Learners should recognize handoff points to data engineering, governance, privacy, legal and domain specialists.

### 6.5 Do not convert vacancy evidence into an employment promise

The corpus supports curriculum relevance, not a guarantee of job readiness, employment, promotion, salary, certification recognition or business results. Roles vary by seniority, domain, country and prior experience. Completion of a course cannot substitute for employer assessment, authorized professional judgment or jurisdiction-specific requirements.

## 7. Responsible-AI interpretation

Two of 126 accepted records (1.6%) explicitly mention AI-assisted analytical work in the retained observable skill fields. One refers to AI-assisted analytics; another to AI-assisted insight generation. The count is reproducible from the accepted corpus, but it is particularly sensitive to the minimal-evidence design: a vacancy may use AI without stating it in the excerpted fields, and a general AI reference may not represent a core responsibility.

The responsible conclusion is narrow. AI literacy is useful as an assisting practice, but the corpus does not justify presenting generative AI as a replacement for SQL, data quality, statistical reasoning, visualization or stakeholder communication. Appropriate learning uses include refining a question, drafting alternative queries, proposing anomaly hypotheses, critiquing a chart, checking whether a narrative omits a limitation, or comparing two supplied outputs.

Human controls remain essential:

- do not place identifiable, confidential, privileged, customer, employee, health, financial or credential data into an unapproved AI service;
- treat generated calculations, code, citations and explanations as unverified drafts;
- reperform material calculations and test queries against known cases;
- trace every retained factual claim to accepted evidence;
- document where AI assistance affected the work;
- keep consequential decisions with an authorized human;
- distinguish model suggestions from observed evidence and professional judgment.

The report itself contains no invented model output and makes no claim that AI use causes better analytical results.

## 8. Legal, credential and portfolio interpretation

The completed rights review classifies the subject as teachable with mandatory controls. Original text, a fully synthetic dataset, compact tables, formulas, SQL examples, accessible chart specifications and original templates can support the learning outcomes without reproducing protected curricula, proprietary datasets, screenshots, exam questions or dashboards.

The exact proposed course title is also used by BCS for a separate exam-based certificate. Any future MTF course must display MTF Institute adjacent to the title and state that its award is a non-degree course-completion certificate, not a BCS, Microsoft, Google, Tableau, IBM or other vendor qualification. No alignment, equivalence, preparation, recognition, partnership or endorsement may be implied.

MTF also has a closely named Udemy course. The completed portfolio review therefore requires a materially original direct-course architecture, case, artifacts, explanations and capstone. The vacancy findings favor a distinct text-first, tool-agnostic, quality-checked decision-support product. They do not authorize reuse of Udemy lectures, projects, transcripts, downloadable files or quizzes.

## 9. Limitations

1. **Purposive sampling.** The 126 vacancies were selected for suitability and variation, not drawn through probability sampling. Percentages describe this corpus only.
2. **Platform bias.** All records were discoverable through public ATS or career pages. Organizations recruiting through private networks, local boards or non-indexed systems are underrepresented.
3. **Geographic imbalance.** The United States and India together account for 70 vacancies, or 55.6%. Country shares cannot be interpreted as comparative demand.
4. **Language and indexing bias.** English-language role discovery favors multinational and digitally mature employers, even though several non-English or multilingual postings were included.
5. **Point-in-time status.** Public application evidence was observed on 31 August 2026. Vacancy status can change immediately after retrieval.
6. **Minimal excerpts.** Short excerpts reduce copyright and replication risk but can under-code duties described elsewhere on the page.
7. **Multi-label coding.** Codes overlap, so technical and responsibility counts cannot be summed into mutually exclusive categories.
8. **Tool-name ambiguity.** A named product can be listed as required, preferred or one option among alternatives. The code records observability, not proficiency level or exclusivity.
9. **Role heterogeneity.** The corpus combines entry, mid-level, senior and hands-on supervisory roles across business intelligence, product, marketing, finance and operations contexts.
10. **No compensation or outcome analysis.** Salary, hiring volume, applicant demand, time-to-hire, retention, promotion and performance outcomes were outside scope.
11. **No causal inference.** A frequent code does not prove that the capability causes hiring, performance or career advancement.
12. **Rights classification gaps.** Dataset A did not carry the same row-level rights-class field as Dataset B, although its method records employer-controlled public ATS sourcing.
13. **AI under-observation.** The 1.6% explicit AI figure reflects retained text fields and cannot estimate actual workplace adoption.
14. **Legal screen, not legal opinion.** The title and rights review did not perform jurisdiction-by-jurisdiction trademark clearance.

## 10. Conclusion

Across 126 accepted current public vacancies, practical data analysis is most consistently visible as the production of trusted reporting from queried and quality-checked data, followed by clear communication to stakeholders. Dashboard/reporting work appears in 77.0% of the corpus, SQL in 69.8% and insight communication in 58.7%. Quality, diagnostic analysis, KPI definition, requirements discovery and documentation form the supporting professional system.

The central curriculum implication is straightforward: teach analysts to move from a bounded question to a reviewable decision-support package. Software fluency matters, but tools should serve an evidence chain that includes data definitions, validation, analytical reasoning, visual communication, limitations and responsible handoff. That is the common practical core supported by this corpus. It is also the boundary that prevents the course from drifting into data engineering, machine-learning specialization, vendor certification or unsupported employment promises.

Readers who need to deepen the adjacent controls around data ownership, access, quality, stewardship and responsible AI can continue with MTF Institute's existing [Data Governance & AI Readiness for Business Professionals](https://mtfinstitute.com/programs/data-governance-ai-readiness-business-professionals/) programme.

## Selected source links

The complete source ledger contains all 126 URLs. The following links illustrate the geographic, industry and source-family variation; they are examples, not an additional sample:

- [Target — Data Analyst, Bangalore](https://target.wd5.myworkdayjobs.com/en-US/targetcareers/job/Data-Analyst_R0000431066)
- [M-KOPA — Business Intelligence Analyst, Gauteng](https://jobs.ashbyhq.com/M-KOPA/df01aeb7-74e4-42ae-9cd3-f5fb79188835)
- [Netflix — Data Analyst, EMEA Production Finance, London](https://netflix.wd1.myworkdayjobs.com/en-US/Netflix/job/Data-Analyst--EMEA-Production-Finance_JR37106)
- [Gecina — Data Analyst internship, Paris](https://gecina.wd3.myworkdayjobs.com/en-US/GecinaExterne/job/Data-Analyst--H-F----Stage_DP1234)
- [Techem — Data Analyst, Eschborn](https://techem.wd3.myworkdayjobs.com/en-US/TechemGermanyExternalCareerSiteAllJobs/job/Data-Analyst-Voll--Teilzeit--m-w-d-_R-104306)
- [AFRY — Data Analyst, Commercial Excellence, Manila](https://jobs.smartrecruiters.com/AFRY/744000111767565-data-analyst-commercial-excellence)
- [Red Bull Australia — Business Intelligence Analyst](https://jobs.smartrecruiters.com/RedBull/744000125258539-business-intelligence-analyst)
- [Modern Woodmen of America — Data Analyst I](https://recruiting.ultipro.com/MOD1003MWA/JobBoard/8990acfc-04f3-d65f-0565-a772e1a0157b/OpportunityDetail?opportunityId=70700f10-7f5f-46b4-b6d7-900cc26964ad)
- [Unipol Assicurazioni — Data Analyst, Italy](https://hdix.fa.em3.oraclecloud.com/hcmUI/CandidateExperience/it/sites/CX_1/jobs?keyword=CLAIMS&mode=location&selectedFlexFieldsFacets=%2525252522AttributeChar6%252525257CCustom_1%2525252522)
- [Cloudflare — Data Analyst](https://job-boards.greenhouse.io/cloudflare/jobs/6955911)
- [GoTo Group — Data Analyst](https://jobs.lever.co/GoToGroup/0eba2e93-3dbf-4479-a0ad-cfc35613beaf)
- [O*NET OnLine — Data Scientists](https://www.onetonline.org/link/summary/15-2051.00)
- [O*NET OnLine — Operations Research Analysts](https://www.onetonline.org/link/details/15-2031.00)
- [US Bureau of Labor Statistics — Data Scientists](https://www.bls.gov/ooh/math/data-scientists.htm)
- [MTF Institute programs catalogue](https://mtfinstitute.com/programs/)
- [BCS — Professional Certificate in Data Analysis](https://www.bcs.org/qualifications-and-certifications/certifications-for-professionals/business-analysis/professional-certificate-in-data-analysis/)



## Citation

When citing or summarizing this material, link to the canonical HTML page: https://mtfinstitute.com/insights/data-analysis-work-126-vacancies-2026/