Data analysis and graph skills

Introduction to Data Analysis and Graph Skills

In Physics, we don't just collect numbers; we look for patterns. Data analysis is the process of turning a table of measurements into a story about how the universe works. Whether you are measuring the speed of a falling ball or the discharge of a capacitor, graphs are your most powerful tool to visualize these relationships. In this chapter, we will learn how to present data professionally and extract hidden information like the gradient and intercept.

Note: This chapter links closely to "Limitation of physical measurements." If you need a refresher on how to calculate uncertainties, check that chapter first!

1. Drawing the Perfect Graph

A good graph should be clear enough for anyone to understand without asking you questions. Here is the step-by-step "Physics Standard" for plotting:

Choosing Your Axes

The independent variable (the one you change) usually goes on the x-axis. The dependent variable (the one you measure) goes on the y-axis. Always label your axes with the quantity and the unit, separated by a forward slash, for example: \(Time / s\) or \(Potential\ difference / V\).

Selecting a Scale

Don't be afraid to use the whole page! Your points should cover at least half of the graph paper provided. Pro tip: Use "sensible" scales. Stick to multiples of 1, 2, or 5. Avoid using 3, 7, or 9 as your scale intervals—they are a nightmare to plot and often lead to silly mistakes!

Plotting Points and Error Bars

Use a sharp pencil to draw small crosses \( \times \). If you have calculated the uncertainty for your measurements, you must draw error bars. These are lines extending above and below (or left and right) of your point to show the range where the true value likely lies.

Quick Summary:
• Axes: Label with "Quantity / Unit".
• Scale: Use 1, 2, or 5; cover >50% of the page.
• Points: Small crosses, not big dots!

2. The Line of Best Fit (LOBF)

Nature is messy, and your points might not line up perfectly. Your Line of Best Fit should be a single, smooth thin line that follows the trend of the data. You want a balanced number of points on either side of the line.

Direct Proportionality vs. Linear Variation

Students often mix these up, but there is a specific difference in Physics exams:
• Directly Proportional: The graph is a straight line AND it passes through the origin \((0,0)\).
• Linear Relationship: The graph is a straight line, but it does NOT have to pass through the origin.

Did you know? If you are asked to "verify a relationship," drawing a graph and showing it is a straight line through the origin is often the best evidence you can provide!

3. Gradient and Intercepts

Most Physics relationships can be rearranged into the form of a straight-line equation:
\(y = mx + c\)

Where:
• \(y\) = the vertical axis variable.
• \(x\) = the horizontal axis variable.
• \(m\) = the gradient (steepness).
• \(c\) = the y-intercept (where the line crosses the y-axis).

Calculating the Gradient

To find the gradient \(m\), always use the Triangle Method:
1. Pick two points on your line of best fit (not your raw data points!) that are as far apart as possible.
2. Calculate the change in \(y\) (\(\Delta y\)) and the change in \(x\) (\(\Delta x\)).
3. Use the formula: \(m = \frac{\Delta y}{\Delta x}\).

The Intercept and "False Origins"

The y-intercept is the value of \(y\) when \(x = 0\). Common Mistake: If your x-axis doesn't start at zero (a "false origin"), you cannot just read the intercept off the y-axis! Instead, pick a point \((x, y)\) on your line and use the formula \(c = y - mx\) to calculate it.

Key Takeaway: The gradient and intercept often represent physical constants. For example, in a graph of \(v\) against \(t\), the gradient is acceleration \(a\).

4. Uncertainties in Graphs

When you have error bars, your gradient isn't a single "perfect" number; it has an uncertainty too. To find this, you may be asked to draw a Worst Acceptable Line.

• This is the steepest (or shallowest) line that still passes through all of your error bars.
• Uncertainty in gradient = \(|best\ gradient - worst\ gradient|\).
• Uncertainty in intercept = \(|best\ intercept - worst\ intercept|\).

5. Linearizing Non-Linear Equations

Sometimes the relationship isn't a simple straight line. We can "force" it into a linear shape to make it easier to analyze.

Standard Linearization (AS Level)

If you have an equation like \(P = IV + I^2R\), and you want to find \(R\), you might plot \(\frac{P}{I}\) on the y-axis and \(I\) on the x-axis.
Rearranging gives: \(\frac{P}{I} = RI + V\).
Now it matches \(y = mx + c\), where the gradient is \(R\)!

Logarithmic Graphs (A2 Level Only)

Note: This section is for International A-level students (Unit 5).

If the relationship is a power law (\(y = kx^n\)):
Take logs of both sides: \(\log(y) = n \log(x) + \log(k)\).
Plot \(\log(y)\) against \(\log(x)\). The gradient is the power \(n\).

If the relationship is exponential (\(y = ke^{nx}\)):
Take natural logs (\(\ln\)): \(\ln(y) = nx + \ln(k)\).
Plot \(\ln(y)\) against \(x\). The gradient is the constant \(n\).

6. Areas and Tangents

Sometimes the information we want isn't the gradient, but the area under the curve or the rate of change at a point.

Areas Under Graphs

The area often represents a physical quantity. For example, the area under a Force-Displacement graph represents Work Done.
• For straight lines, use geometry (triangles and rectangles).
• For curves, you must "count the squares." Calculate the "value" of one square (width \(\times\) height in their respective units) and multiply by the total number of squares.

Tangents to Curves

If a graph is curved, the gradient is changing. To find the gradient at one specific moment (the instantaneous rate of change):
1. Use a ruler to draw a straight line that just touches the curve at that point.
2. Ensure the "angles" between the curve and the ruler are equal on both sides.
3. Calculate the gradient of that straight line.

Summary Checklist

• Significant Figures: Ensure your calculated gradient has the same number of significant figures as your raw data.
• Units: Did you include units for your gradient? (e.g., if \(y\) is in \(m\) and \(x\) is in \(s\), the gradient is in \(m/s\)).
• Ratios: You can use ratios of data points to test if a power relationship exists (e.g., if doubling \(x\) always quadruples \(y\), then \(y \propto x^2\)).
• Command Words: If the exam says "Determine," you usually need to use your graph. If it says "Show that," you must show every step of your calculation clearly.