Overview

Row

$6.74B

$1,046.96K

45

+7.8%

Store 20

Store 33

Store 35

Sales are stable, not trending — a steady baseline with sharp, predictable seasonal spikes rather than sustained growth or decline across 2010–2012.
Dec is the seasonal peak month, with the largest average weekly sales of the year.
Holiday weeks carry a statistically significant lift of roughly 7.8% over non-holiday weeks (Welch t-test, p < 0.01).
Store performance is highly uneven — a wide gap separates the top store (Store 20) from the lowest-volume store (Store 33), pointing to structural rather than purely operational causes.
Macroeconomic indicators (temperature, fuel price, CPI, unemployment) show negligible correlation with weekly sales in this dataset — seasonal and holiday timing are far stronger, more actionable signals.

Align inventory and staffing with the seasonal calendar, concentrating resources ahead of the November–December peak.
Model holiday weeks as a distinct forecasting category with a dedicated uplift factor rather than the general weekly average.
Investigate root causes at bottom-performing stores — store size, local market density, and competition — before applying a chain-wide fix.
Tighten forecasting cycles for high-volatility stores, where week-to-week swings are largest relative to their own baseline.
De-prioritize macroeconomic indicators in short-term forecasting, given their negligible correlation with sales; revisit for longer-horizon models only.
Extend this work with a formal time-series model (ETS, ARIMA, or Prophet) per store, using the seasonal and holiday signal identified here.

Data source: Walmart Sales dataset (Kaggle) · Period: February 2010 – October 2012 · Dashboard built in R with flexdashboard, plotly, and DT