Sharing & Publishing

Getting the Insight to the People Who Decide

Analysis that no one sees changes nothing. Sharing and publishing is the final step: distributing the report and its insights to the right audiences, in the right format, through the right channel — and doing it reliably and repeatedly. This is also where governance matters most: the moment data leaves the analyst's machine, access, accuracy, and privacy become everyone's problem.

Publishing pushes one governed source of truth out to dashboards, portals, and stakeholders.

Channels for Sharing

• Live dashboards: Power BI, Tableau, or Looker — self-serve, always current, best for ongoing monitoring.
• Data portals / workspaces: a shared, governed location where teams find the trusted version.
• Static artifacts: PDF decks, exported tables, and documents for point-in-time communication.
• Notebooks: a reproducible end-to-end story for technical audiences — like the one at the end of this series.

Make It Reproducible, Not Manual

The difference between a one-off chart and a real deliverable is repeatability. Export from code so the same report can be regenerated on demand; schedule refreshes so the numbers are never stale; and version the output so people know which copy is current. A pipeline that rebuilds the report on a schedule beats a spreadsheet someone updates by hand.

In Python — export the deliverables

Python · pandas

import pandas as pd

# The summary table from the Reporting step
summary = (orders.groupby(["region", "category"])
                 .agg(sales=("sales", "sum"), profit=("profit", "sum"))
                 .reset_index())

# Publish to several formats from a single source of truth
summary.to_csv("retail_summary.csv", index=False)        # for analysts
summary.to_excel("retail_summary.xlsx", index=False)     # for business users

# A styled HTML table to embed in a portal or email
(summary.style
        .format({"sales": "${:,.0f}", "profit": "${:,.0f}"})
        .background_gradient(subset=["sales"], cmap="Blues")
        .to_html("retail_summary.html"))

In SQL — publish to a shareable view

SQL

-- A governed view is the warehouse-native way to "publish":
-- one definition, granted to the right roles, queried by every BI tool.
CREATE OR REPLACE VIEW reporting.retail_summary AS
SELECT region, category,
       SUM(sales)  AS sales,
       SUM(profit) AS profit
FROM orders
GROUP BY region, category;

GRANT SELECT ON reporting.retail_summary TO ROLE bi_readers;

Tools Commonly Used

• Python: Pandas to_csv/to_excel/to_html, Styler; nbconvert for notebooks.
• SQL: governed VIEWs and role-based GRANTs as the published interface.
• BI & collaboration: Power BI / Tableau publishing, scheduled refreshes, shared workspaces.

Best Practices

• Publish a single source of truth; avoid emailing copies that immediately drift out of date.
• Automate generation and refresh so the report is always current and reproducible.
• Apply access control and respect privacy/compliance (e.g., GDPR) before anything is shared.
• Match the format to the audience — a dashboard for monitoring, a deck for a decision meeting.

Example: Publishing the Retail Report

For the retail dataset, publishing might involve:

• A scheduled Power BI dashboard for the regional managers to monitor sales and margin.
• A governed warehouse view that every BI tool reads from, so the numbers always agree.
• A reproducible notebook — exactly the one below — that walks the full pipeline for technical reviewers.