2.2 Histograms, Frequency Polygons, and Time Series Graphs

Histograms and Frequency Polygons

Histograms and frequency polygons are tools for visualizing how data is distributed. They help you quickly spot patterns like where data clusters, how spread out it is, and whether the distribution is symmetric or skewed. Frequency polygons are especially useful when you need to compare two or more datasets on the same graph.

Time Series Graphs

Time series graphs track how a variable changes over time. They're essential for spotting long-term trends, seasonal patterns, and unusual spikes or dips in data. You'll see these used everywhere from economics to climate science.

Histograms

Construction of Histograms

A histogram uses bars to show how frequently data values fall within specific ranges (called intervals or bins). Unlike a bar graph, the bars in a histogram touch each other because the data is continuous or ordered along a number line.

Here's how to build one:

1. Identify the data type. Histograms work with quantitative data, which can be continuous (height, weight, temperature) or discrete (number of siblings, number of pets).

2. Find the range. Subtract the minimum value from the maximum value in your dataset.

3. Choose the number of intervals and calculate the width. A good rule of thumb is 5 to 20 intervals, depending on your sample size. Calculate the interval width:

$\text{Interval width} = \frac{\text{maximum value} - \text{minimum value}}{\text{number of intervals}}$

All intervals must be the same width, and they should not overlap.

1. Set a starting point. For continuous data, start at a value less than or equal to the minimum (like rounding down to 0, 10, or 100). For discrete data, start at a value the data can actually take on.

2. Count frequencies. Tally how many data values fall into each interval. You can use raw frequency (counts) or relative frequency (the proportion of total data in each interval).

3. Draw the histogram. Place intervals on the x-axis and frequency (or relative frequency) on the y-axis. Each bar's height represents how many values fall in that interval. Bars should be adjacent with no gaps between them.

Once built, a histogram lets you quickly see the shape of your distribution: Is it symmetric? Skewed left or right? Are there gaps or clusters?

Frequency Polygons for Data Comparison

A frequency polygon is a line graph version of a histogram. Its main advantage is that you can overlay multiple distributions on the same graph, which makes comparison much easier than placing histograms side by side.

To construct one:

1. Build a histogram (or at least determine the intervals and frequencies) for each dataset. Use the same interval width and starting point across all datasets.
2. Find the midpoint of each interval. For example, if an interval runs from 10 to 20, the midpoint is 15.
3. Plot a point at each midpoint, with the height equal to the frequency (or relative frequency) of that interval.
4. Connect the points with straight line segments.
5. If you're comparing multiple datasets, use different colors or line styles and include a legend so the reader can tell them apart.

Time Series Graphs

Analysis of Time Series Graphs

A time series graph plots data values in chronological order, with time on the x-axis and the variable of interest on the y-axis. These graphs are built for answering one question: How does this variable change over time?

When analyzing a time series graph, look for three things:

• Overall trend. Is the data generally increasing, decreasing, or staying flat over the full time period? For example, global average temperatures show an increasing trend over the past century, while certain manufacturing costs have shown a decreasing trend.
• Cyclical patterns or seasonality. Some data repeats in predictable ways. Seasonality refers to patterns that repeat within a fixed period, like retail sales spiking every December. Cyclical patterns repeat over longer, less predictable periods, like economic expansion and recession cycles.
• Outliers or unusual observations. Look for data points that break sharply from the overall pattern. A sudden spike in hospital visits during a flu outbreak, for instance, would stand out on a time series graph.

Context matters here. A sharp drop in air travel in early 2020 makes sense once you know about the pandemic. Always consider external factors (policy changes, natural disasters, new technology) that might explain what you see in the data.

Data Interpretation and Representation

Statistical graphics give you a visual way to analyze quantitative data. Whether you're using a histogram, frequency polygon, or time series graph, the goal is the same: make patterns, trends, and outliers visible so you can draw meaningful conclusions.

Keep in mind that every graph has limitations. The number of intervals you choose for a histogram can change how the distribution appears. A time series graph might hide short-term variation if the time intervals are too wide. Effective interpretation means understanding both what the graph shows and what choices went into building it.