Store dashboard: DAX used, visualization, validation & findings

Live Power BI dashboard

Published report walmart_visualization — Executive, Store, Product, and Demand pages on Gold aggregates. Open in Power BI ↗

Analysis pages: 08 Executive · 09 Store · 10 Product · 11 Demand · Overview

Part A — DAX measures used on Store (Page 2)

Full 33-measure catalogue lives on 08 Executive. Store page uses these measures (paste-ready):

/* Cards + bar + summary */
Total Units =
SUM ( 'agg_daily_store_sales'[units_sold] )

Total Sales Value =
SUM ( 'agg_daily_store_sales'[sales_value] )

Avg Sell Price =
AVERAGE ( 'agg_daily_store_sales'[avg_sell_price] )

/* Matrix — category share (format %) */
Category Share =
AVERAGE ( 'agg_store_category_sales'[category_sales_share] )

Line chart on this page typically uses the column agg_daily_store_sales[units_sold] (Sum) with Legend store_id — not a separate measure.

Part B — Dashboard 2 screenshot (updated)

Power BI page STORE after layout polish: year + store slicers, state donut, daily units-by-store line, category share matrix (no bogus 33% column), store summary table, and store ranking bar. Sales card still needs tighter currency formatting (see §4).

1. Scope & method

2. Visualization & interaction

2.1 Layout

┌──────────────────────────────────────────────────────────────┐
│ [ Year slicer ]              [ Total Units ] [ Sales Value ] │
│ [ store_id slicer ]  [ State donut ]  [ Units by date×store ]│
│ [ Category share matrix ] [ Store summary ] [ Store ranking ]│
└──────────────────────────────────────────────────────────────┘

2.2 Visual-to-field map

2.3 Interaction design

2.4 Suggested click-path

1. Set year to a full calendar year (2015) for fair store ranks.
2. Read donut — confirm CA ≈ 45% of sales.
3. Note CA_3 vs CA_4 gap on the ranking bar.
4. Click CA on the donut → inspect CA-only line trends and matrix.
5. Clear → select CA_3 and CA_4 in store slicer → compare mix and price.
6. Return to all stores → read FOODS share stability in the matrix.

3. Key pattern findings

3.1 California leads; CA_3 is the flagship

Within California, CA_3 alone is 38% of CA sales and 17.1% of network sales ($32.70M). CA_4 is last network-wide ($12.47M). Ratio CA_3 / CA_4 = 2.62×. California contains both the best and worst stores — state label alone does not explain performance.

3.2 Store ranking (matches dashboard bar)

*Mean of daily avg_sell_price on agg_daily_store_sales.

3.3 Category mix is FOODS-led in every store

Mean category sales share across stores: FOODS 58.1% · HOUSEHOLD 29.5% · HOBBIES 12.4%. Extremes:

• Highest FOODS: WI_2 64.5%, WI_3 63.5%
• Lowest FOODS / highest HOUSEHOLD: CA_2 (FOODS 49.3%, HOUSEHOLD 38.4%)
• Highest HOBBIES: WI_1 15.7%, CA_4 15.3%

Assortment differs by store, but FOODS remains the primary revenue engine everywhere — consistent with Executive category columns.

3.4 Prices are tightly clustered

Store-level average sell prices sit in a narrow band $4.34–$4.45 (network ~$4.39). CA_3’s leadership is volume, not a higher price point. TX stores skew slightly cheaper; WI_1 is the dearest on this metric.

3.5 Partial 2016 coverage still matters for KPIs

The units line is now at day grain (good). KPI cards and ranked totals with all years selected still include only 143 days in 2016 (through 22 May). Network average daily units rise from ~37.8K (2015) to ~41.8K (2016 YTD) — so filter to a full year when comparing store ranks for stakeholders.

4. Dashboard quality notes

• ✓ Category share matrix: bogus per-row Total column at 33.33% removed. Total row now shows overall mix (~FOODS 58% · HOBBIES 12% · HOUSEHOLD 30%), consistent with Gold.
• ✓ Units line uses continuous date (“Sum of units_sold by date and store_id”), not year roll-up.
• ! Total Sales Value card still shows long decimals ($191.57754604M). Match Executive: Currency + display units Millions, 2 decimal places → $191.58M.
• ! Line chart density: all 10 stores on one daily chart is hard to read. Use the store slicer (multi-select ≤3) or Small multiples by store_id.
• ✓ Donut + ranked bar + matrix + summary table remain a solid store-ops layout; year + store slicers support focused comparison.

5. Recommendations

1. Benchmark ops and assortment against CA_3; deep-dive CA_4 and high-HOUSEHOLD CA_2.
2. Format the Store page sales card like Executive ($191.58M).
3. Limit the daily line to ≤3 stores via slicer (Multi-select with CTRL = Off) or use small multiples.
4. Use full-year filters (2012–2015) when ranking stores for stakeholders.
5. Keep price monitoring light — dispersion is small; volume and mix drive gaps.
6. Sync year slicer with Executive; optionally sync store_id when drilling from Page 1.

6. Reproduce

python - <<'PY'
import pandas as pd
from pathlib import Path
P = Path('data/gold_export')
s = pd.read_parquet(P/'agg_daily_store_sales.parquet')
print(s.groupby('state_id')['sales_value'].sum())
print(s.groupby('store_id')['sales_value'].sum().sort_values(ascending=False))
print(pd.read_parquet(P/'agg_store_category_sales.parquet')
      .pivot_table(index='store_id', columns='cat_id', values='category_sales_share'))
PY

Screenshot: assets/store-dashboard-page2.png. Related: 08 Executive.