Job search
How to Share a Power BI Portfolio with Recruiters: Public Reports, Files, and Access Limits
Share a Power BI portfolio with an authorized audience. Compare public reports, account access, PBIX files and a readable, explicitly bounded backup.
Share a Power BI portfolio through a route the recruiter can actually use: a public report built from material you may expose, a permissioned report for a known account, or an inspectable file with a readable backup. Decide the audience before generating a link. State the data, licensing requirements and interaction limits, then verify the delivered version in its intended access context.
LandOffer publishes this guide and offers job-search tools. We checked Microsoft's public documentation and created a local, fictional workshop-capacity backup for Rosalind. Its PDF and arithmetic were inspected; it was authored locally, not exported from Power BI. No Power BI account, public embed, recipient invitation or PBIX file was tested. Use recent roles on LandOffer to identify whether the reader needs to inspect reporting logic, visual decisions or both.

Decide what you are allowed to expose
A portfolio link has two separate jobs: deliver the work and preserve its permitted audience. Begin with the report's inputs, model and pages. Candidate ownership of a screenshot does not establish permission to release the employer's underlying records. A sample made independently from invented data can offer a clearer sharing boundary than a lightly disguised workplace workbook.
Rosalind's fictional employer report contains internal workshop bookings. She has no permission to publish it. The completed delivery decision excludes that workbook, its screenshots and its data model. She instead creates six original, synthetic session records for a separate educational example. The resulting scope is narrow: explain a capacity metric and present the six observations clearly.
That decision leaves some work unshown. Her sample cannot establish that she implemented an employer's refresh pipeline or improved workshop attendance. Those claims would need their own authorized evidence. It can show how she defines a metric, checks a total and prepares a reader-facing explanation. A focused contribution can still be useful.
Record excluded material before selecting the delivery route. Include private source files, hidden pages, attachments and any details that would reveal an individual or client. If a source's reuse terms are unclear, choose different inputs or resolve the permission question first. A disclaimer beside an exposed report cannot retract the exposure.
Understand what Publish to web makes public
Microsoft's Publish to web documentation describes viewing without authentication and warns that underlying detail-level model data may be accessible even when the report does not display it. A public link therefore needs a review of the model as well as the visible chart. Treat the route as a decision to release the permitted material to a broad audience.
The guide also describes author licensing, editing and resharing requirements, tenant administration and unsupported report configurations. A report that looks fine in the author's workspace may still be ineligible for this publishing route. Row-level security reports and several connection or visual arrangements have restrictions. Check the current limitations against the actual model before promising a public version.
Do not ask an employer to disable a tenant control for a portfolio. If the organization's policy prevents public publishing, choose an authorized alternative. For independently created sample work, the author still needs to establish the account, tenant and model conditions. An invented dataset does not automatically make the technical publishing operation available.
Record the public version's date and contents. Microsoft's guide discusses caching and delays in visible updates. A corrected author view does not prove the public reader has the correction. Recheck the actual delivered version after a change, and know where its embed code is managed. No embed code was created or deleted for Rosalind's exercise.
Separate named-user access from a public link
A permissioned report is a different delivery arrangement. Microsoft's report-sharing guide describes recipient identity, permissions, licenses and capacity exceptions. The practical question is whether this specific reader can open this specific report with the intended account. “I copied the URL” is not an answer to that question.
An existing-access link does not grant access. Organization-scoped links and external-recipient sharing also have different requirements. For an external reader, check the supported guest route, account identity and administrator settings. Keep read access separate from permissions that allow building on the model or resharing content. Hiding a field in a report is not a substitute for an access control.
Avoid the blanket claim that every recruiter can view a report for free. The documented capacity exceptions depend on the qualifying arrangement, including where the report and its model reside. The default license requirements and exceptions need to be checked together. The guide's description is not proof that a candidate's workspace meets them.
For Rosalind, named-user sharing remains unverified because there is no recipient account or hosted report. Her delivery record says so directly. If a recruiter later requests an account-specific review, she would verify that route before replacing the backup with a gated link. She would not describe the documentation review as a successful invitation test.
Decide whether a PBIX file helps the reader
A PBIX file can expose a different level of work from a browser report. Before offering one, inventory what the actual file contains and what it requires to open and use. Check the data, queries, model, connections and any references to private locations. Sharing a file does not excuse releasing material that was excluded from the public report.
Microsoft's Power BI Desktop installation guide describes a Windows environment and current supported versions. A recruiter using an ordinary browser should not be assumed to have a compatible Desktop installation. Ask whether the requested review includes opening the authoring file, and provide a readable explanation that does not depend on that installation.
Distinguish opening the saved file from refreshing its sources. A recipient may need additional access or configuration for a refresh even if the file opens. Specify which behavior you have checked. If the exercise depends on local sample inputs, retain those inputs and instructions in the authorized package. Do not imply that a refresh was tested merely because a PDF exists.
Rosalind's record marks PBIX as “not created or opened.” This is a completed decision about the present packet, not a claim about a file's quality. Her current deliverable demonstrates a metric explanation and static presentation. A future modeling sample would require a separately inspected PBIX and an honest account of its opening and refresh conditions.
Prepare a static backup with a defined view
A PDF gives the reader a stable view of selected pages. It cannot preserve every interaction. Microsoft's PDF export guide explains different export choices and limitations. In the service, default values and current values can produce different views. Hidden-page choices, unsupported visuals and other report features also matter. Inspect the result instead of treating export completion as visual approval.
If a displayed view matters to the claim, record its filters and selected period. Check every exported page for readable labels, complete tables, clipping and a clear metric definition. Confirm that the file's contents match the accompanying sentence. A static page showing one selection should not be presented as evidence that all drill-through states were reviewed.
The guide also describes sensitivity-label and encryption behavior. A downloaded PDF can remain restricted. Verify whether the recipient needs a compatible viewer or credentials, and do not remove an organization protection to make a job-application attachment easier. A file extension alone does not establish accessibility.
A short recording can demonstrate an interaction, but it needs the same authorized-content boundary and a useful static explanation. Name the starting state and the steps shown. Avoid narration that claims a tested model refresh, accessible public link or recipient permission when the recording only shows the author's screen. Keep its purpose tied to a specific review task.
Inspect Rosalind's completed backup
Our original PDF contains a six-session table, a total and a bounded interpretation. All counts are invented. A row represents a workshop session, with registered seats and available seats. The ratio is capacity utilization across those sessions; it is not attendance, satisfaction or a count of unique participants.
| Session | Registered seats | Available seats |
|---|---|---|
| January | 12 | 20 |
| February | 18 | 20 |
| March | 14 | 20 |
| April | 21 | 24 |
| May | 22 | 24 |
| June | 15 | 24 |
| Total | 102 | 132 |
The calculation is 102 divided by 132, or 77.27% when rounded to two decimal places. Summing the counts before dividing answers the stated aggregate-capacity question. An unweighted average of session percentages would answer a different question when capacities vary. The PDF keeps the numerator and denominator beside the percentage.
We rendered and inspected the complete one-page PDF, checked its extracted text and verified the retained arithmetic. It includes the fictional-data disclosure and says it was authored locally rather than exported from Power BI. The full-page rendering, extracted text and retained calculation support the backup's local readability and stated numbers. They do not measure Power BI's export fidelity, a recruiter's viewer behavior or a hosted report's access.

The retained local packet includes the synthetic CSV, PDF, delivery record and explanatory README. The PDF's interpretation stops at summarizing registered capacity. It does not recommend adding sessions or claim that a real organization acted. The data analyst resume guide can help translate this kind of bounded contribution into a concise application account.
Keep a completed delivery record
Use the record to distinguish a chosen route from a verified route. Rosalind's completed record contains the following decisions, including the work still required before a hosted handoff.
| Item | Recorded decision | Verified scope |
|---|---|---|
| Employer workbook | Excluded | No permission to expose it in the fictional case |
| Sample inputs | Six invented sessions | Local CSV and totals checked |
| Public report | Possible future route | No account, tenant or model eligibility test |
| Named-user report | Possible requested review | No recipient identity or license check |
| PBIX | Absent from current packet | No opening or refresh claim |
| Readable backup | Original one-page PDF | Full local rendering and text inspected |
The unsent handoff wording is specific: “Independent workshop-capacity exercise using six synthetic sessions. The readable backup defines registered seats divided by available seats and reports 102/132, or 77.27%. It was authored locally; no public Power BI link or PBIX is included.” That sentence tells the reader what exists without inventing a recipient test.
For an actual hosted report, replace the unverified entries only after the relevant checks. Record the report URL, permitted audience, account used for the check, date, starting view and backup filename. Revisit those entries when you alter permissions or content. A successful check on an earlier version should not silently certify a later one.
Match the handoff to the application
Follow the employer's requested format and attachment limits. If the application asks for a portfolio URL, provide the authorized, tested starting point and label the useful backup in its accompanying description. If it asks for a file, verify the actual selected attachment before submitting. A polished report can still arrive through the wrong field or with the wrong version attached.
Make the deliverable's scope easy to inspect. Lead with the question, your contribution, the permitted data source and the view the reader should examine. Explain any login or software requirement before the reader encounters it. If the requested route cannot be provided, describe the available deliverable accurately rather than claiming equivalent interaction.
Choose a reporting responsibility from recent roles on LandOffer, then deliver the authorized report or backup that demonstrates it.


