Data Visualization: All About 2D Graphs

Subscribe to RichardOnData here: https://www.youtube.com/channel/UCKPyg5gsnt6h0aA8EBw3i6A?sub_confirmation=1 Doing data visualization appropriately is important if you're going to do it at all. As a general principle, SIMPLER graphs are much, much better for stakeholders than are more COMPLEX graphs. This becomes doubly true when we invoke effects like 3D when it is not necessary. Here are some examples of terrible graphs: https://www.businessinsider.com/the-27-worst-charts-of-all-time-2013-6 Here are the "good" types of 2D graphs: * Bar Charts/Column Charts: Used for summarizing counts or percentages of categorical data. To summarize another variable these can be put side-to-side or occasionally stacked. You should generally lean towards a bar chart if you have long labels or a lot of values. Likewise you should lean towards a column chart if you want to illustrate particular large values or sorted data, or you have sequential groups. * Histograms: Used for summarizing counts or percentages of quantitative/continuous data. These can also be smoothed to create what's called a density plot. * Box Plots: Used for summarizing distributions of quantitative/continuous data. It's very convenient to summarize multiple groups side-by-side as well to make comparisons. These provide the minimum, first quartile, median, third quartile, and maximum of a distribution. * Scatterplots: Used to investigate the relationship between two quantitative/continuous variables. One convenient extension is to create a lot of these, if you have many variables, so you can quickly detect meaningful relationships, multicollinearity, and do appropriate feature selection for a machine learning problem. * Line Plots: Used to investigate a trend in a quantitative/continuous variable over time. It's generally okay to put a few variables on the same plot. * Pie Charts: Used to summarize counts or percentages of categorical data. These accentuate the part-to-whole relationship. However, they are not preferable because your eyes cannot easily detect relative differences between multidimensional objects. #DataVisualization #DataScience #DataViz PayPal: [email protected] Patreon: https://www.patreon.com/richardondata BTC: 3LM5d1vibhp1F7pcxAFX8Ys1DM6XLUoNVL ETH: 0x3CfC599C4c1040963B644780a0E62d45999bE9D8 LTC: MH8yPjvSmKvpmRRmufofjRB9hnRAFHfx32