More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Data Story Studio Pro

    1

    Create distinctive, traceable data stories with 24 scenario playbooks, 18 visual directions, and English or Chinese outputs.

    $2.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Data Story Studio Pro

    Data Story Studio Pro

    Example session with this skill installed

    Create a complete English product-conversion story from this fictional January 2026 cohort, observed through 28 February 2026. Use a dark technical art direction and four complementary analytical chapters. Return the complete readable report inline in Markdown.

    Metric contract: distinct people from one tracked cohort. Each later stage is a subset of the preceding stage; stage populations overlap and must never be added together.

    Source IDStageDistinct people
    visitVisited12,000
    signupSigned up3,600
    activateActivated2,100
    paidPaid840

    Show original-cohort shares, adjacent conversion with its own denominator, mutually exclusive dispositions, and the final paid/complement partition. Include formulas and source IDs. Recommend an evidence-review priority without claiming causation or guaranteed improvement. Label it as a fictional local creator demonstration; buyer-Agent/Agensi MCP execution remains unverified.

    • Read your context and instructions
    • Compiled the data story studio

    From first visit to paid use

    Finding: 840 of 12,000 people reach paid use:

    7% of the original visitor cohort. The largest recorded count difference is between visiting and signing up. This identifies a boundary to investigate, rather than a cause of non-conversion.

    Scope: One fictional January 2026 cohort, observed through 28 February 2026. Each successive stage is a subset of the preceding stage. Counts are distinct within each stage; the stages overlap in people.

    1. A cohort through the product

    The four stages describe one population, not four separate populations. Keep the original 12,000-person denominator visible when asking how much of that cohort reaches each stage.

    StageRecorded peopleShare of original cohortSource IDs
    Visited12,000100%visit
    Signed up3,60030%visit, signup
    Activated2,10017.5%visit, activate
    Paid8407%visit, paid

    Formula: Stage share = stage count / initial cohort × 100%. The stage counts must not be added: doing so would count some people repeatedly.

    2. The conversion at each boundary

    Adjacent conversion uses the preceding stage, whereas the first chapter used the original cohort. These are different denominators answering different questions.

    BoundaryExact ratioConversionCount differenceSource IDs
    Visited → Signed up3,600 / 12,00030%8,400visit, signup
    Signed up → Activated2,100 / 3,60058.3%1,500signup, activate
    Activated → Paid840 / 2,10040%1,260activate, paid

    Formula: Adjacent conversion = next stage / preceding stage × 100%. Displayed rates are rounded after calculation; 2,100 / 3,600 is exactly 7/12, rather than exactly 58.3%. A zero preceding count would make its ratio undefined.

    3. Where the original cohort is accounted for

    Subtracting successive nested counts produces mutually exclusive dispositions. These can be reconciled to the original population; the overlapping source stages cannot.

    Recorded dispositionPeopleOriginal-cohort shareSource IDs
    Visited, not recorded as signed up8,40070%visit, signup
    Signed up, not recorded as activated1,50012.5%signup, activate, visit
    Activated, not recorded as paid1,26010.5%activate, paid, visit
    Recorded as paid8407%visit, paid

    Formula: Boundary count = preceding count − next count. Reconciliation: 8,400 + 1,500 + 1,260 + 840 =

    12,000. The disposition shares sum to 100%.

    The 8,400-person visit-to-signup difference is the largest absolute boundary difference. It does not reveal motives, identify a failed feature or establish that those people will never convert.

    4. The final outcome, once

    Outcome at the observation cutoffPeopleOriginal-cohort shareSource IDs
    Reached paid use8407%visit, paid
    Not recorded at the paid stage11,16093%visit, paid

    Formulas: Paid share = 840 / 12,000 × 100% = 7%. Complement = (12,000 − 840) / 12,000 × 100% = 93%. The final partition reconciles to 840 + 11,160 =

    12,000.

    Next review: Audit event coverage, identity matching and the observation window around Visited → Signed up. Investigate the experience with source records before choosing an intervention. This is a descriptive review priority, not evidence of causation or guaranteed improvement.

    Example status: Fictional data; local creator demonstration. Buyer-Agent and Agensi MCP execution have not been independently verified. The deliverable shown here is the complete inline report.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Give your spreadsheet a story people can follow. Data Story Studio Pro guides your AI Agent from source inspection and metric definitions to a distinctive visual report with traceable findings, precise charts and an editable handoff. Built for client work, team reviews, research communication and public explanations.

    What makes it useful

    • Evidence before headlines: define population, period, units, comparison and numerator/denominator; distinguish zero, missing and unavailable values.
    • Complete analytical stories: choose complementary chapters that answer different questions, with source-linked findings, exact evidence tables, formulas and a useful next decision.
    • Distinctive visual authorship: 18 visual-direction guides cover composition, typography, chart treatment, annotation and interaction. Create a new direction when the brief calls for it.
    • Coherent edits and exports: keep current findings, marks, tables and translated copy synchronized when source values or filters change.
    • English and Chinese: English instructions support English, Chinese or a requested bilingual edition from one consistent data model.

    24 scenario playbooks

    Growth and product: sales contribution, acquisition efficiency, gross margin, market comparison, activation funnels, cohort retention, survey preferences and customer journeys.

    Operations and finance: monthly operating reviews, cash/budget bridges, project calendars, response times, delivery reliability, manufacturing quality and inventory age.

    Research and public communication: environmental records, energy use, public-service access, nonprofit participation/outcomes and research sample coverage.

    Learning and systems: learning progression, aggregate hiring pipelines, event participation, and qualitative knowledge/process explanations. Combine or extend these playbooks for a new brief.

    Design choice

    Use visual guides such as Swiss Index, Night Observatory, Editorial Broadsheet, Field Atlas, Bauhaus Blocks, Calibration Plate or Evidence Folio—or develop an original direction. Carry its typography and structure through the analytical chapters and evidence, with purposeful responsive behavior.

    Requested outputs

    Interactive HTML, editable SVG, PNG sharing graphics, PDF reports and companion CSV/JSON can be authored when your Agent environment supports the format. Inputs can include CSV, XLSX, JSON, verified tables and source material. XLSX requires an available spreadsheet reader; the optional inspection helper uses openpyxl. No paid provider, seller backend or creator demonstration runtime is required.

    What the examples represent

    The six gallery images are actual local creator outputs using fictional data. Our separate bilingual showcase contains 24 case families and 84 complementary analytical chapters. The purchase provides reusable instructions and reference guidance; the scenario/style catalogs are not a bundle of prebuilt reports, datasets, fonts or demonstration assets.

    Execution and delivery

    Skill access follows Agensi’s agent connection. Generated reports and their original editable sources are separate from protected Skill instructions. Buyer execution and Agensi MCP multi-file delivery have not been verified; outputs depend on the Agent’s available tools. Local HTML reopening was checked through HTTP. The demonstrated PDFs are raster reports with non-searchable text; retain SVG and data for editing. Every real assignment needs source and rendered-output verification.

    Foundation provenance

    The MIT-0 chart-generator and data-analyzer instruction files were starting inspiration. Narrative methods, metric contracts, scenario guidance, art direction and acceptance guidance here are newly authored and do not require those Skills or copy their executable templates.

    Version 0.3.0.

    Agent tools and file-delivery capabilities determine available outputs. Buyer execution, Agensi MCP multi-file/reference delivery and cross-provider compatibility remain untested; the local creator showcase is separate. The 24 playbooks and 18 design guides are not prebuilt report assets. Sources and metric contracts require per-project verification. The demonstrated PDF route is raster-based with non-searchable text. No downloadable/offline Skill access, platform hosting, population validity or guaranteed business outcome is promised.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Browser

    Read user-authorized sources, calculate in an available interpreter, author original report files, and inspect the rendered output in a browser. No account credentials, environment secrets, seller backend or paid provider are required. External data or asset fetching is optional and needs the user’s permitted sources; protected Skill resources must follow Agensi’s execution rules.

    Listed1 month ago
    Updated14 days ago

    What's inside

    Frequently Asked Questions

    Reusable English instructions, 13 reference guides and an optional source-inspection helper, accessed under Agensi’s current agent-connection rules. The 24 scenario playbooks and 18 directions are design guidance. Gallery reports, sample datasets, fonts and the separate creator showcase are not a prebuilt template bundle included with the purchase.

    Yes. Select or combine playbooks for growth, research, operations, finance, learning and qualitative explanations. Use a visual-direction guide or author a new direction for the reader and question. Three or four chapters are appropriate when they answer complementary questions; page count follows the brief and evidence.

    The workflow defines units, population, period, numerator/denominator and comparison before calculation; retains source IDs and formulas; reconciles applicable components; and checks current charts, tables and exports. It guides verification rather than guaranteeing arbitrary source correctness, causal inference or scientific validity.

    Instructions are entirely English. Generated work can be English, Chinese or a requested bilingual edition using one consistent data model, separate text dictionaries and available CJK font fallback. Review the actual rendered edition for your target environment.

    Request interactive HTML, editable SVG, PNG, PDF and CSV/JSON where your Agent has permitted analysis, rendering and file tools. XLSX needs an available reader; the optional Python helper uses openpyxl. The demonstrated PDF is raster-based and non-searchable. No paid provider or seller backend is mandatory. Buyer/MCP multi-file delivery and universal cross-provider execution are not verified.

    New in Research & Analysis