5:03
How To Choose The Right Graph (Types of Graphs and When To Use Them)
UNDATABLE
Overview
This video explains the five most common types of graphs and when to use them effectively. It covers bar charts (vertical, horizontal, stacked), histograms, pie charts, scatter plots, and line charts. The presenter emphasizes choosing the right graph for the data type and the message you want to convey, offering practical tips and examples for each. The goal is to help viewers create clearer, more persuasive data visualizations by sticking to these fundamental chart types and understanding their strengths.
How was this?
Save this permanently with flashcards, quizzes, and AI chat
Chapters
- The video focuses on the five most popular and useful chart types for data visualization.
- Understanding these charts helps in communicating data effectively and persuasively.
- The goal is to equip viewers with the knowledge to select the appropriate chart for their specific needs.
Knowing which chart to use is crucial for accurately representing data and ensuring your audience understands the intended message, preventing misinterpretation.
The presenter introduces the idea that mastering a few key chart types is more beneficial than knowing many obscure ones.
- Vertical bar charts are ideal for comparing distinct categories, especially when there are fewer than seven.
- Horizontal bar charts are better for ranking when categories exceed seven, or for communicating performance and election results.
- Stacked bar charts show how different parts contribute to a whole, but should be kept simple to avoid confusion.
Bar charts offer clear visual comparisons, making it easy to identify differences, rankings, and proportional contributions within a dataset.
A vertical bar chart showing app downloads in 2020, highlighting that TikTok, not Zoom, was the most downloaded app. A stacked bar chart breaking down US downloads for TikTok and Zoom to show a smaller difference within a key market.
- Histograms, similar to bar charts, display the distribution of data over a continuous variable (e.g., age, income, time).
- They are less about individual data points and more about revealing broader patterns and trends.
- Histograms help in understanding the shape of a distribution, such as population age demographics.
Histograms are essential for understanding the frequency and spread of data within a continuous range, revealing underlying patterns that might be missed otherwise.
A histogram showing income distribution in the US, or population distribution by age, demonstrating how patterns can change when looking at different regions (e.g., Nigeria's younger population).
- Pie charts are often criticized for being ineffective at comparing multiple data points due to the difficulty in judging angles and areas.
- If used, limit pie charts to a maximum of five data points.
- Tips for effective pie charts include ranking segments, placing labels directly next to segments, and using color strategically.
While visually appealing, pie charts can be misleading. Understanding their limitations and best practices ensures they are used only when appropriate and with maximum clarity.
The presenter challenges viewers to identify the third most downloaded app from a pie chart, then shows the same data in a bar chart, demonstrating the bar chart's superior clarity for comparison.
- Scatter plots are used to visualize the relationship or correlation between two continuous variables.
- They can show positive correlation, negative correlation, no correlation, identify clusters, and spot outliers.
- Scatter plots are highly persuasive for demonstrating a relationship between two factors.
Scatter plots are powerful tools for exploring potential connections between variables, helping to identify trends, relationships, and anomalies in data.
The presenter explains that scatter plots are ideal for showing how one variable might influence another, or if they move independently.
- Line charts are the best choice for illustrating how a variable changes over a period of time.
- They are effective for tracking trends, such as stock prices, website traffic, or search interest.
- Line charts can reveal seasonality or cyclical patterns in data.
Line charts provide a clear visual narrative of change, making it easy to follow trends and understand the dynamics of a variable over time.
A line chart showing the search volume for the term 'line chart' over several years, illustrating its increasing popularity with seasonal dips.
- Stick to the five discussed chart types unless you are an expert data visualizer.
- Focus on clear communication and persuasive storytelling with your chosen chart.
- Consider learning about the 'data-ink ratio' to maximize the impact of your visualizations.
Applying these principles ensures your data visualizations are not only accurate but also effectively communicate insights and persuade your audience.
The presenter suggests learning about the data-ink ratio as a way to improve chart effectiveness.
Key takeaways
- Choose a graph type that directly matches the type of data you have and the message you want to convey.
- Bar charts are versatile for comparisons and rankings, but be mindful of the number of categories.
- Histograms reveal the distribution and patterns within continuous data, not individual categories.
- Use pie charts sparingly and with strict adherence to best practices due to their limitations in comparison.
- Scatter plots are essential for identifying and illustrating relationships or correlations between two variables.
- Line charts are the standard for showing trends and changes over time.
- Prioritize clarity and effective communication over using complex or numerous chart types.
Key terms
Bar ChartVertical Bar ChartHorizontal Bar ChartStacked Bar ChartHistogramDistributionContinuous VariablePie ChartScatter PlotCorrelationOutlierLine ChartData-Ink Ratio
Test your understanding
- What is the primary difference between a bar chart and a histogram?
- Under what circumstances should you choose a horizontal bar chart over a vertical one?
- Why are pie charts often discouraged by data visualization experts, and what are the best practices if you must use one?
- How can a scatter plot help in understanding the relationship between two different metrics?
- When is a line chart the most appropriate choice for visualizing data?