Project Documentation

Sales Report Dashboard

An interactive property sales report built in Power BI and rebuilt for the web with ECharts and Leaflet

Version 1.0 · November 2026 · Portfolio project · Anonymized data

Executive summary

Sales Report Dashboard is an interactive property sales report. It was designed in Power BI for stakeholders who need a quick view of commercial performance (where sales come from, which business units drive them, and where cancellations happen) and was rebuilt as a static web page so anyone can explore it without a Power BI licence.

87,073sales-order lines
5Power BI pages rebuilt
7cross-linked filters
45cities on the map
⚠️ All names and amounts in the demo are anonymized; the figures are realistic in shape but not real.
Dashboard overview page
Overview page (light mode)

Open the live demo →

Features

Global filters

Province, City, SBU, Division, Cluster, Product type and SO year range. The City list narrows to the selected province.

Click to cross-filter

Clicking a bar or a map bubble applies that value as a filter, like cross-filtering in Power BI.

Sales map

Bubble map of net sales by city; bubble area is proportional to net sales.

Ranked comparisons

Net sales by province, top 10 cities, SBU and product type (with share of total).

Cancellation trend

Cancelled orders per calendar month, split by the four main product types plus Other.

Summary & detail tables

An expandable province → city summary, and a sortable, searchable, paginated detail table.

Light & dark mode

Follows the system setting and the portfolio's theme toggle; chart colours are re-validated for each mode.

Responsive

Works on phones: filters and charts stack, and axis numbers hide when space is tight.

Privacy by design

No real names or amounts are published; the original files stay in a private repository.

Report pages: Power BI → web

Each Power BI page and visual has a web counterpart:

Power BI pagePower BI visualWeb dashboard
CoverTitle, logo, Last Refreshed cardPage header with data cut-off date (no company logo)
OverviewSlicers (7)Sticky filter bar shared by all tabs
OverviewAzure Map (Province, City)Leaflet bubble map on OpenStreetMap
OverviewTop 10 bar charts (Province, City, SBU)ECharts horizontal bars with value labels
OverviewDonut: net sales by product typeRanked bars with % share (easier to compare close values)
OverviewLine: Cancellation Province by monthLine chart by product type
SummaryMatrix (Province → City → SBU → Division → Cluster)Expandable Province → City table with totals
DetailTableSortable, searchable, paginated table
Detail | Drill ThroughDrill-through pageReplaced by click-to-filter + the Detail tab

Metric definitions

MetricDefinitionSource column
Net sales (excl. VAT)Sum of net sales excluding VAT. Cancellation lines are negative, so this is already net of cancellations.SalesNetExPPNAmt
Net sales (incl. VAT)Same as above, including VAT (PPN).SalesNettInPPNAmt
Net units soldSum of unit quantity: +1 for a sale line, −1 for a cancellation line.UnitQty
CancellationsNumber of sales-order lines whose order type is a cancellation (ZC01 / ZC02). Includes cancel-and-rebook addenda.OrderType → IsCancel
Year filterYear of the first sales-order date.SOFirstDate
Cancellation monthCalendar month of the sales-order date.MonthSODateNumber

Amounts are shown in rupiah with T = trillion, B = billion, M = million.

Architecture

Private · Power BIPreparation · PythonPublic · GitHub PagesSource data (Excel)87,073 SO lines · 67 columnsPower BI report (.pbix)5 pages · slicers · DAX measuresSelect & clean19 columns + 3 helper columnsAnonymizegeneric names · scaled amounts + noiseAggregate → data.json57,603 grouped rows · dictionary-encodedDashboard (index.html)filters · ECharts · Leaflet · tablesVisitor's browserall filtering runs client-side
The original data and .pbix stay in a private repository; only anonymized, aggregated data is published.

No backend

One HTML file plus data.json. Aggregation for every filter combination happens in the browser in milliseconds.

Compact data

Text values are stored once in dictionaries and rows reference them by index: 2.2 MB raw, about 0.5 MB gzipped.

Libraries

Apache ECharts 5.5 for charts, Leaflet 1.9 with OpenStreetMap tiles for the map, both loaded from a CDN.

Data & anonymization

The source is a sales-order view exported to Excel. Before anything was published, every identifying value was replaced:

FieldOriginalPublished
CompanyReal legal entity namesCompany A, B, C…
SBUBusiness-unit names incl. locationGeneric: Residential Core, Commercial Core, Regional East…
DivisionReal division namesDivision A, B, C…
Cluster (373)Real project / cluster namesCluster 001, 002…
AmountsReal rupiah values× hidden random factor, plus ±15% noise per row
Order typeZS01, ZC01, …Removed; only the derived IsCancel flag is kept
Province, city, dates, product type—Kept (not identifying)
The relative patterns (rankings, seasonality, product mix) remain realistic, so the dashboard still tells a credible story, but no figure can be traced back to the real business.

How to use

1

Start with the Overview

The KPI tiles show the totals for the current filters; the map and bars show where sales come from.

2

Narrow down

Pick values in the filter bar, or click a bar or bubble. Press Reset to clear everything.

3

Compare regions

Open Summary and click a province to see its cities side by side.

4

Find specific rows

Open Detail, search by city, company or cluster, and click a column header to sort.

Tip: tabs can be linked directly, e.g. …/#summary or …/#detail.

Run locally

The page loads data.json with fetch, so serve the folder over HTTP rather than opening the file directly:

git clone https://github.com/diimasogan/PowerBI-SalesReport-Dashboard.git
cd PowerBI-SalesReport-Dashboard
python3 -m http.server 8000
# open http://localhost:8000

To rebuild this documentation after editing the text:

python3 -I docs/_source/build_docs.py

Limitations & notes