10 · Project — Multi-Source Executive Dashboard¶
This capstone combines every Level 2 module into one deliverable: a
multi-source executive dashboard over Orders and Targets, published and
interactive.
1. Inputs¶
Orders (8 rows, Level 1 Module 1) and Targets (Level 2 Module 7):
| Region | Sales | Target | Attainment |
|---|---|---|---|
| East | 2210 | 2000 | 110.5% |
| West | 3750 | 3500 | 107.1% |
| Central | 920 | 1000 | 92.0% |
Attainment computed as Sales/Target: 2210/2000=1.105, 3750/3500=1.0714, 920/1000=0.92 — confirmed against Level 2 Module 7's Section 3.
2. Data layer¶
- Combine
OrdersandTargetsvia a cross-database join (Module 7) onRegion, rather than a blend, so a single calculated field can compute Attainment without cross-source limitations. - Add
Pct of Region SalesandDiscount Tierfrom Level 2 Module 1 for drill-down detail: Furniture's East share example (Bookcase 950/2210 = 43.0%) and the Tier A/B/C/Review classification per order. - Extract the combined source (Module 8) with aggregation enabled for the
Region/Category-level sheets, and a separate non-aggregated extract (or
the same extract without the aggregation option) for the row-level
detail sheet — verify both report
SUM(Sales) = 6880before proceeding.
3. Sheets to build¶
- KPI header: three floating text/number objects — Total Sales ($6,880), Total Profit ($866: 180+90+18+96+330-40+42+150), Overall Attainment (6880/6500 = 105.8%, using Target total 2000+3500+1000=6500).
- Region bar chart:
SUM(Sales)by Region with a reference line at each Region's Target (dual-axis or a Gantt-bar target overlay) — West's bar (3750) should visibly clear its 3500 reference line; Central's bar (920) should fall short of its 1000 line. - Category breakdown: Furniture 4050, Electronics 2650, Office Supplies 180 — built as a second sheet, linked to the Region chart via a filter action (Level 2 Module 2).
- Discount Tier detail table: row-level
Order ID,Product,Discount Tierfrom Level 2 Module 1's classification (Tier A: 1001, 1005; Tier B: 1008; Review: 1006; Tier C: 1002, 1003, 1004, 1007) — kept inside a Show/Hide container so it doesn't compete with the KPI header.
4. Interactivity checklist¶
- Clicking a Region bar filters the Category sheet — verify clicking "East" shows Furniture (1200+950=2150) and Office Supplies (60), summing to 2210.
- A Show/Hide toggle reveals the Discount Tier detail table.
- A phone device layout stacks the KPI header above a single combined chart rather than the 3-column desktop arrangement.
5. Formatting and storytelling pass (Module 5)¶
- Title reads the finding:
"West leads at $3,750 (107% of target) — Central trails at 92%". - Diverging color on the Attainment field, centered at 100%: West and East render in the "above target" hue, Central in the "below target" hue — correctly flagging Central as the one region under its goal.
6. Publish and verify¶
- Publish to Tableau Server/Cloud (Module 9) as a separate data source plus a workbook referencing it, with "Download Full Data" restricted for viewer-level permissions.
- Attach a nightly extract refresh schedule.
- Post-publish check: open the published URL and re-confirm all headline numbers — Total Sales $6,880, Total Profit $866, Overall Attainment 105.8%, and each Region's bar/target relationship — exactly match the hand-computed values in Sections 1–3. Any mismatch means the publish used a stale or mis-joined data source and should be re-published after fixing the join.
How It Actually Works¶
This capstone's headline numbers depend on several mechanisms from earlier Level 2 modules composing correctly — verifying it end-to-end means checking each mechanism did what it should, not just that the final numbers look plausible:
- The cross-database join (Section 2.1) means Attainment is computed
inside a single VizQL query against the joined result —
SELECT Region, SUM(Sales)/SUM(Target) FROM (Orders JOIN Targets ON Region) GROUP BY Regionconceptually — rather than two independently-rendered sheets whose numbers are only visually adjacent (Level 2 Module 7, Section 4). This is exactly why 2210/2000=1.105 can be trusted as a single computed field rather than something a viewer has to mentally divide across two charts. - Extract aggregation (Section 2.3) intentionally splits into two
physical extracts here because the Region/Category sheets and the
row-level Discount Tier table need different grains — an aggregated
extract collapsing to Region/Category totals cannot serve the
Discount Tier detail table's per-Order
IFlogic (Level 2 Module 1), since that row-level information no longer exists in an aggregated.hyperfile. Verifying both extracts independently reportSUM(Sales) = 6880(Section 2.3) confirms the aggregation step didn't silently drop or double-count rows during that collapse. - Overall Attainment's two computation paths (this module's Exercise)
are a direct test of aggregation order:
SUM(Sales)/SUM(Target)computes one ratio from two grand totals (6880/6500 = 1.0585), while the Sales-weighted average of per-region ratios (Σ(Sales_i × Attainment_i) / ΣSales_i) mathematically reduces to the exact same expression once expanded —Σ(Sales_i × Sales_i/Target_i)isn't generally equal toΣSales_i × ΣSales_i/ΣTarget_i, so the fact that this dataset's numbers come out matching (~105.8% either way) is a coincidence of this specific data, not a general algebraic identity — worth flagging explicitly, since students who assume the two methods are always equivalent for any dataset will get burned on a differently-shaped one. - The Show/Hide detail table and the filter action share the same underlying data, so clicking "East" on the Region chart and separately expanding the Discount Tier table should show mutually consistent detail — Furniture+Office Supplies (2150+60=2210) on one sheet and exactly the East-region Order IDs (1001, 1003, 1006) on the other — any divergence between them indicates the filter action's field binding or the extract's grain doesn't match across the two sheets.
Cheat sheet — capstone checklist¶
| Layer | Delivered |
|---|---|
| Data | Orders + Targets cross-database join, verified totals |
| KPIs | Total Sales, Total Profit, Overall Attainment |
| Charts | Region-vs-Target bars, Category breakdown, Tier detail |
| Interactivity | Filter action, Show/Hide container, phone layout |
| Formatting | Finding-based title, diverging Attainment color |
| Publish | Server publish, restricted download, refresh schedule |
Exercise¶
Compute Overall Attainment a second, independent way — as the
Sales-weighted average of the three regional attainments, i.e.
(2210×1.105 + 3750×1.0714 + 920×0.92) / 6880 — and confirm it still comes
out to approximately 105.8%, matching the Total Sales ÷ Total Target method
used in Section 3.