Time Series Line Charts

Overview

Line charts display variables over time, showing trends, patterns, and temporal relationships. They're the most common chart type for time series data in MMM.

Purpose: Visualize how variables change over time, identify trends, seasonality, and relationships between time series.

When to Use Line Charts

Best For:

• Showing trends over time

• Identifying seasonal patterns

• Comparing multiple variables temporally

• Visualizing time series data

• Spotting outliers or anomalies

Examples:

• Marketing spend trends over months

• KPI performance over time

• Comparing TV vs. Digital spend

• Viewing adstock effects

Creating a Line Chart

Step 1: Select Variables

• Choose 1-8 variables (2-5 recommended)

• Can include KPI for comparison

• Search to filter if many variables

Step 2: Choose Line Chart Type

• Click "Line Chart" button

• Icon: 📈 Trending up

Step 3: Configure Options

• Dual Y-Axis: Use if variables have different scales

• Include KPI: Add target variable to chart

Step 4: Generate

• Click "Generate Chart"

• Chart appears in right panel

Chart Elements

X-Axis (Time):

• Date labels

• Monthly intervals by default

• Adjusts to data frequency

Y-Axis (Value):

• Variable values

• Auto-scaled to data range

• Left axis for all variables (or first half if dual axis)

Lines:

• Each variable is a different colored line

• 2-pixel width

• Markers at data points

Legend:

• Shows variable names and colors

• Click to hide/show series

• Bottom or side position

Dual Y-Axis

When to Use:

• Variables with very different scales

• Example: Spend ($1000s) vs. CTR (0-5%)

• Prevents smaller values from being flattened

How It Works:

• First half of selected variables → Left Y-axis

• Second half → Right Y-axis (opposite side)

• Each axis scales independently

Example:

Variables selected: [TV_Spend, Digital_Spend, CTR, Conversion_Rate]

Left Y-Axis: TV_Spend, Digital_Spend ($thousands)
Right Y-Axis: CTR, Conversion_Rate (percentages)

Reading Line Charts

Trends:

• Upward slope = increasing

• Downward slope = decreasing

• Flat = stable

Seasonality:

• Regular up-and-down patterns

• Annual, quarterly, or monthly cycles

• Peaks at same times each year

Correlations:

• Lines moving together = positive correlation

• Lines moving opposite = negative correlation

• No pattern = little correlation

Outliers:

• Sharp spikes or dips

• Points far from trend

• May indicate data issues or special events

Common Patterns

Growing Trend:

• Consistent upward movement

• Marketing increasing over time

• Business growth

Declining Trend:

• Consistent downward movement

• Channel being phased out

• Market contraction

Seasonal Pattern:

• Regular cyclical movements

• Holiday peaks

• Summer dips

Volatile:

• Erratic up and down

• Campaign-driven spend

• Promotional effects

Stable:

• Relatively flat line

• Consistent always-on spend

• Baseline performance

Interactive Features

Zoom:

• Mouse wheel to zoom in/out

• Selection box (click and drag)

• Focus on specific time periods

Pan:

• Drag left/right when zoomed

• Navigate across timeline

Tooltip:

• Hover over line for exact values

• Shows date and value

• Multiple series shown

Legend Click:

• Hide/show individual lines

• Isolate specific variables

• Compare subsets

Reset:

• Toolbar button

• Return to full view

• Clear all zooms

Use Cases

Trend Analysis

Identify direction:

Question: Is Digital spend increasing?
Chart: Line chart of Digital_Spend over time
Observation: Steady upward trend
Conclusion: Yes, Digital growing consistently

Seasonal Detection

Find patterns:

Question: Are there seasonal patterns in sales?
Chart: Line chart of KPI
Observation: Regular Q4 peaks every year
Conclusion: Strong holiday seasonality

Variable Comparison

Compare channels:

Question: How does TV compare to Digital?
Chart: Line chart with both variables
Observation: TV stable, Digital growing
Conclusion: Shift from TV to Digital over time

Correlation Check

Validate relationships:

Question: Does TV spend drive sales?
Chart: Line chart of TV_Spend and KPI
Observation: Lines move together, KPI lags slightly
Conclusion: Positive correlation with lag effect

Best Practices

Variable Count:

• 2-3 variables: Easy to read

• 4-5 variables: Still manageable

• 6-8 variables: Gets cluttered

• 8 variables: Use multiple charts

Scale Considerations:

• Similar scales: Single Y-axis

• Different scales: Dual Y-axis

• Mix types: Consider separate charts

Time Range:

• Full history for trends

• Zoom for detailed periods

• Compare same periods year-over-year

Color Choice:

• Automatic colors usually fine

• Distinct colors help differentiation

• Consider colorblind-friendly palettes

Troubleshooting

Lines are flat/compressed: → Use dual Y-axis if scales very different

Too cluttered: → Reduce number of variables → Use legend to hide some series → Create multiple charts

Missing data points: → Check data upload → Verify variable exists in all periods → Review data quality

Can't see pattern: → Zoom into specific period → Check if seasonal (need full year) → Try different variables

Examples

Example 1: Marketing Spend Trends

Variables: TV_Spend, Digital_Spend, Radio_Spend Observation: Digital increasing, TV stable, Radio declining Insight: Budget shifting to digital

Example 2: KPI with Seasonality

Variables: Sales_KPI Observation: Regular Q4 peaks, Q1 dips Insight: Strong holiday seasonality, plan budgets accordingly

Example 3: Spend vs. Performance

Variables: Total_Spend, KPI Configuration: Dual Y-axis Observation: Both trending up, KPI faster than spend Insight: Improving efficiency over time

Next Steps

After creating line charts:

• Use Scatter Plots to explore specific relationships

• Check Correlation Matrix for all variable pairs

• Apply insights in Model Building

• Compare with Decomposition results after modeling