Practical “which Power BI visual to use when” guide - Part 1

 

1) First ask: What question are you answering? (Fast decision guide)

✅ A) Compare values across categories

➡️ Use Bar/Column chart
Examples:

• “Which product sold the most?”
• “Which region has highest revenue?”

✅ B) Trend over time (change, seasonality)

➡️ Use Line / Area / Combo chart
Examples:

• “Monthly revenue trend last 24 months”
• “Tickets trend day-by-day”

✅ C) Part-to-whole (composition)

➡️ Use Treemap / 100% stacked bar (avoid pie unless very few categories)
Examples:

• “Revenue contribution by product category”
• “Spend by cost head”

✅ D) Relationship / correlation

➡️ Use Scatter/Bubble chart
Examples:

• “Does discount % increase sales?”
• “Does delivery time impact customer rating?”

✅ E) Distribution (how values spread)

➡️ Use Histogram (custom) / Box plot (custom) / Column with bins
Examples:

• “Order value distribution”
• “Response time distribution (SLA)”

✅ F) Progress vs Target / KPI tracking

➡️ Use KPI / Card / Gauge (limited use) / Bullet chart (best if available)
Examples:

• “Sales vs Target”
• “SLA compliance %”

✅ G) Flow / stages (pipeline)

➡️ Use Funnel (only when it truly narrows stage-by-stage)
Examples:

• “Leads → Qualified → Proposal → Won”

✅ H) Contribution to change

➡️ Use Waterfall chart
Examples:

• “Why profit changed from last month to this month?”

✅ I) Ranking changes over time

➡️ Use Ribbon chart
Examples:

• “Top 5 products rank by month”

✅ J) Detailed reporting / audit view

➡️ Use Table / Matrix
Examples:

• “Invoice-level detail”
• “P&L with hierarchy (GL → Sub-GL → Account)”

✅ K) Geography-based insights

➡️ Use Map / Filled map (only when geography matters)
Examples:

• “Sales by state/district”
• “Service tickets by city”

---

2) Visuals explained with when to use + real case

2.1 Bar / Column (Most used)

Best for: Comparing categories; ranking.
When to choose:

• More than 5–7 categories? Prefer bar (horizontal) for readability.
• Need Top N? Use bar + Top N filter.

Real case: Retail Sales Dashboard

• Visual: Bar chart
• Question: “Top 10 SKUs by Revenue”
• Slice by: Region, Channel, Month
  ✅ Clear ranking, easy to compare.

Avoid: Too many categories without Top N (becomes clutter).

---

2.2 Line chart (Trend over time)

Best for: Time series (daily/weekly/monthly).
Real case: Support Ops

• Visual: Line
• Metric: Tickets created vs resolved per day
• Insight: Identify backlog growth or improvements.

Best practice: Add moving average or trend line for noisy data.

---

2.3 Combo chart (Line + Column)

Best for: Two metrics with different behavior (e.g., volume + rate).
Real case: Sales Performance

• Columns: Monthly Sales
• Line: Gross Margin %
  This helps answer: “Sales increased but did margin drop?”

---

2.4 Stacked / 100% Stacked

Best for: Composition across categories (and change over time).

• Stacked bar: shows absolute + composition
• 100% stacked: shows share % only

Real case: HR Analytics

• 100% stacked by Month: Attrition split (Voluntary / Involuntary)
  Shows “attrition mix changing?” irrespective of total.

---

2.5 Treemap (Part-to-whole with many categories)

Best for: Many categories where you want quick proportional view.
Real case: IT Spend

• Treemap: Cost by Vendor → Service Type
  Instantly shows biggest spend blocks.

Avoid: Precise comparison (humans struggle comparing rectangle areas).

---

2.6 Pie/Donut (Use carefully)

Use only if:

• Max 2–5 categories
• One category clearly dominates
  Real case: Payment Method Share (UPI, Card, COD, Wallet) if only 4.

Avoid: 10 slices = unreadable. Use bar instead.

---

2.7 Waterfall (Explaining variance)

Best for: “What changed and why?”
Real case: Finance

• Profit change from Q1 to Q2 explained by: Price effect, Volume effect, Cost increase, FX impact.
  Perfect for bridge analysis.

---

2.8 Scatter / Bubble (Correlation + outliers)

Best for: Relationship between 2 measures, detect clusters/outliers.
Real case: E-commerce pricing

• X: Discount %
• Y: Units sold
• Bubble size: Revenue
  Helps answer: “Which products respond to discount and which don’t?”

---

2.9 Funnel (Stage drop-off)

Best for: Sequential stages with drop-offs.
Real case: Sales CRM pipeline

• Leads → Qualified → Proposal → Closed
  Shows conversion and leakage.

Avoid: If your process isn’t sequential (then use bar).

---

2.10 KPI / Card / Gauge (Single number)

Card: “Total Sales”, “Total Tickets”, “MTTR”
KPI visual: shows value + target + trend
Gauge: okay for exec view, but limited comparison (bullet is better).

Real case: SLA Monitoring

• KPI: SLA compliance % vs target 95% with trend over time
• Cards: Open tickets, Avg resolution time

---

2.11 Table vs Matrix (Detail & Hierarchies)

Table: flat row-level detail (invoice list, ticket list)
Matrix: hierarchical (Region → Country → City, or GL hierarchy)

Real case: P&L report

• Matrix: Revenue, COGS, Opex with drill-down
• Conditional formatting for variance % (green/red)

---

2.12 Decomposition Tree (Root cause analysis)

Best for: “Why is this high/low?” with drill-based explanation.
Real case: Manufacturing defects

• Start: Total Defects
• Drill: Plant → Line → Shift → Operator → Defect type
  Great for operations and RCA.

---

2.13 Key Influencers (Driver analysis)

Best for: “What factors influence churn / late delivery / defect?”
Real case: Customer churn

• Target: Churn = Yes/No
• Explain by: Tenure, plan type, complaints, region
  Gives data-driven drivers quickly.

---

2.14 Maps (Geography)

Use if location matters and you have geo fields (state/city/lat-long).
Real case: Service Coverage

• Filled map by state: ticket density
• Drill to city hotspots

Avoid: If geography is just a label and doesn’t add insight.

---

3) Common mistakes (and what to do instead)

❌ Using Pie for everything

✅ Use bar/treemap for many categories.

❌ Too many visuals on one page

✅ Use 3–6 strong visuals, then drill-through for details.

❌ No context (no target, no benchmark)

✅ Add target lines, previous period, variance %, tooltips.

❌ Same scale issues in combo charts

✅ Use combo only when it makes sense; ensure right axis is clear.

---

4) Recommended “real dashboard layouts” (templates)

Template A: Sales Performance (Exec)

1. Cards: Revenue, Profit, Margin%, YoY%
2. Line: Revenue trend (Monthly)
3. Bar: Top 10 Products
4. Map: Revenue by state (optional)
5. Waterfall: Profit variance vs last quarter
6. Matrix: Region → Channel breakdown with drill

Template B: Ops / Support

1. KPI: SLA%, MTTR
2. Line: Created vs Resolved tickets
3. Bar: Tickets by Category
4. Scatter: Resolution time vs CSAT (outliers)
5. Table: Ticket-level drill-through

---

5) Quick cheat-sheet (one-liners)

• Compare categories: Bar/Column
• Trend: Line
• Trend + Target: Line + constant line / KPI
• Part-to-whole (many): Treemap / stacked
• Explain change: Waterfall
• Correlation/outliers: Scatter
• Pipeline stages: Funnel
• Geography: Map
• Root cause drill: Decomposition tree
• Drivers: Key influencers
• Details/audit: Table/Matrix