2025 Week 51 | Power BI: Holiday Traffic: Who Came, Who Left 🎄

Introduction

This report visualizes account activity movement over time, not just a net change. For each date, it shows two flows at once:

  • New Active Accounts (inflow) above zero

  • Lost Active Accounts (outflow) below zero

The goal is to make daily shifts easy to read, with a gradient background to emphasize the “above/below” balance.

Requirements

  • A Date table related to the fact table (orders).

  • A clear definition of Active Account using a rolling window (example: last 30 days).

  • Two independent measures (so you can have both + and − on the same date):

    • New Active Accounts: active today, not active yesterday

    • Lost Active Accounts: active yesterday, not active today (plotted as negative)

  • A stacked column chart with:

    • dynamic Y-axis min/max (to keep the visual aligned)

    • two spacer (dummy) measures to “fill” the chart up to the axis limits

  • A background gradient image layered behind the chart for the balance effect.

Dataset

Sales Orders data with at least:

  • Order Date

  • Customer ID
    Plus a connected Date dimension.

Share

After you finish your workout, share on Bluesky or LinkedIn using the hashtags #WOW2025 and #PowerBI, and tag @MMarie, @shan_gsd, @KerryKolosko (on BlueSky).

Solution

  • Measures created:

    • Active Accounts (rolling window)

    • New Active Accounts (today’s active set minus yesterday’s)

    • Lost Active Accounts (yesterday’s active set minus today’s, shown negative)

    • Axis Radius / Axis Min / Axis Max (to control chart range consistently)

    • Growth Spacer / Loss Spacer (technical fillers for the balance layout)

  • Visual built as a stacked column chart with the series ordered like:
    New Active Accounts → Growth Spacer → Lost Active Accounts → Loss Spacer

  • Y-axis range set via measures (fx) so the layout stays stable when filtering.

  • Gradient background added as an image and placed behind the chart; spacers can be styled to blend into the background while the real values remain readable.

Via Data Stories Gallery