Introduction to Data Visualisation

In Physics B (Advancing Physics), graphs are much more than just drawings; they are the primary tools we use to uncover the "stories" hidden within our experimental data. Whether you are measuring the resistance of a wire or the speed of a falling object, a well-drawn graph allows you to spot patterns, identify errors, and calculate physical constants with precision. In this chapter, we will look at how to move from a table of messy numbers to a clear, professional plot that can withstand scientific scrutiny.

Simple Distribution Plots (Dot-plots)

Before we jump into complex graphs, we often start with a dot-plot. This is a simple way to look at a single set of measurements to see how they are "spread out."

Imagine you and five friends all measure the length of the same metal rod. You might get slightly different values: \(10.1 \text{ cm}\), \(10.2 \text{ cm}\), \(10.1 \text{ cm}\), \(10.3 \text{ cm}\), and \(11.5 \text{ cm}\). By plotting these as dots on a single horizontal axis, you can easily see:

  • The Mean or Median: Where the cluster of dots is centered.
  • The Spread: How much the data varies. A wide spread suggests low precision.
  • Outliers (Anomalies): That \(11.5 \text{ cm}\) measurement looks suspicious! It’s far away from the others and might be a mistake that needs investigating.

Plotting Professional Scatter Graphs

Most of your work will involve scatter graphs, where you plot an independent variable (the one you change) against a dependent variable (the one you measure).

1. Choosing Your Axes and Scales

The rules for a "Physics-standard" graph are strict but helpful:

  • Labelling: Always label axes with the quantity / unit. For example: \(V / \text{V}\) or \(t / \text{s}\). The forward slash is the standard convention.
  • Scaling: Use simple intervals (like 1s, 2s, 5s, or 10s). Avoid awkward scales like 3s or 7s that make plotting difficult.
  • Filling the page: Your data points should occupy at least half of the available grid area. Don't squash a beautiful experiment into a tiny corner!

2. Uncertainty (Error) Bars

Data points aren't perfect. We use uncertainty bars to show the "margin of error" for each measurement. If your ruler has an uncertainty of \(\pm 0.1 \text{ cm}\), you draw a small vertical or horizontal line through your point extending \(0.1\) units in both directions. These bars are essential for establishing the validity of your data—if a best-fit line misses the error bars, the relationship might not be what you think it is.

The Line of Best Fit

A best-fit line (or curve) represents the overall trend of the data. It should have a roughly equal number of points above and below the line. Note: A best-fit line does not have to go through the origin unless the physics theory says it must!

Key Terms to Remember:

Linear Relationship: When the data points form a straight line, usually following the equation \(y = mx + c\).
Interpreting the Intercept: The \(y\)-intercept (\(c\)) is the value of \(y\) when \(x = 0\). In some experiments, a non-zero intercept might indicate a systematic error (like a zero error on a meter).

Calculating Gradients and Intercepts

The gradient (slope) of a graph often represents a physical constant, such as the resistance of a component or the acceleration of an object.

Step-by-Step Gradient Calculation:

  1. Choose two points on your line of best fit (not your raw data points!).
  2. Ensure these points are far apart (at least half the length of the line) to reduce percentage error.
  3. Use the formula: \(m = \frac{\Delta y}{\Delta x} = \frac{y_2 - y_1}{x_2 - x_1}\)
  4. Don't forget the units for the gradient! If you plot Force (\(\text{N}\)) against Extension (\(\text{m}\)), your gradient unit will be \(\text{N m}^{-1}\).

Uncertainty in Gradients

In Physics B, we don't just calculate one gradient; we need to know how much we can trust it. We do this by finding the worst-fit lines.

How to find the uncertainty in the gradient:

  1. Draw your "Best" line of best fit.
  2. Draw the "Steepest" possible line that still passes through all your uncertainty bars.
  3. Draw the "Shallowest" possible line that still passes through all your uncertainty bars.
  4. The uncertainty in your gradient is roughly: \(\text{Uncertainty} = \frac{\text{steepest gradient} - \text{shallowest gradient}}{2}\).

Quick Review: The "worst" lines are the ones that are as far away from your best-fit line as possible while still being "valid" (passing through the error bars).

Other Types of Plots

While scatter graphs are the "gold standard," the syllabus mentions a few others for specific uses:

  • Pie-charts: Useful for showing proportions (e.g., different components of energy in a system).
  • Log graphs: These are used to establish the validity of data that covers a huge range of values or follows power-law relationships. You will see these more in Imaging and Signalling to handle large numbers of bits or signal-to-noise ratios.

Common Mistakes to Avoid

1. The "Dot-to-Dot": Never simply connect your data points with jagged lines. Nature usually follows smooth trends!

2. Forcing the Origin: Don't force your line through \((0,0)\) if the points clearly point somewhere else. You might be ignoring a systematic error.

3. Small Triangles: When calculating a gradient, using a tiny triangle leads to massive errors. Always "Go Big" with your gradient triangle.

Summary Key Takeaways

1. Graphs turn numbers into visual evidence.
2. Use Quantity / Unit for all labels.
3. Uncertainty bars show the reliability of your measurements.
4. The gradient uncertainty is found by comparing your best-fit line to the steepest and shallowest possible lines (worst-fit lines).
5. Gradients and intercepts have physical meanings and units.