{"id":123709,"date":"2022-04-08T18:16:17","date_gmt":"2022-04-08T18:16:17","guid":{"rendered":"https:\/\/blog.finxter.com\/?p=292393"},"modified":"2022-04-08T18:16:17","modified_gmt":"2022-04-08T18:16:17","slug":"how-to-color-a-scatter-plot-by-category-using-matplotlib-in-python","status":"publish","type":"post","link":"https:\/\/sickgaming.net\/blog\/2022\/04\/08\/how-to-color-a-scatter-plot-by-category-using-matplotlib-in-python\/","title":{"rendered":"How to Color a Scatter Plot by Category using Matplotlib in Python"},"content":{"rendered":"
Problem Formulation<\/h2>\n
Given three arrays:<\/p>\n
\n
The first two arrays x<\/code> and y<\/code> of length n<\/code> contain the (x_i, y_i)<\/code> data of a 2D coordinate system.<\/li>\n
The third array c<\/code> provides categorical label information so we essentially get n<\/code> data bundles (x_i, y_i, c_i)<\/code> for an arbitrary number of categories c_i<\/code>. <\/li>\n<\/ul>\n
Question<\/strong>: How to plot the data so that (x_i, y_i)<\/code> and (x_j, y_j)<\/code> with the same category c_i == c_j<\/code> have the same color?<\/p>\n
Solution: Use Pandas groupby() and Call plt.plot() Separately for Each Group<\/h2>\n
To plot data by category, you iterate over all groups separately by using the data.groupby()<\/code> operation. For each group, you execute the plt.plot()<\/code> operation to plot only the data in the group.<\/p>\n
In particular, you perform the following steps:<\/p>\n
\n
Use the data.groupby(\"Category\")<\/code> function assuming that data is a Pandas DataFrame containing the x<\/code>, y<\/code>, and category<\/code> columns for n<\/em> data points (rows). <\/li>\n
Iterate over all (name, group)<\/code> tuples in the grouping operation result obtained from step one.<\/li>\n
Use plt.plot(group[\"X\"], group[\"Y\"], marker=\"o\", linestyle=\"\", label=name)<\/code> to plot each group separately using the x<\/code>, y<\/code> data and name<\/code> as a label.<\/li>\n<\/ol>\n
Here’s what that looks like in code:<\/p>\n
import pandas as pd\nimport matplotlib.pyplot as plt # Generate the categorical data\nx = [1, 2, 3, 4, 5, 6]\ny = [42, 41, 40, 39, 38, 37]\nc = ['a', 'b', 'a', 'b', 'b', 'a'] data = pd.DataFrame({\"X\": x, \"Y\": y, \"Category\": c})\nprint(data) # Plot data by category\ngroups = data.groupby(\"Category\")\nfor name, group in groups: plt.plot(group[\"X\"], group[\"Y\"], marker=\"o\", linestyle=\"\", label=name) plt.legend()\nplt.show()<\/pre>\n
![\"\ud83d\udcac\"](\"https:\/\/s.w.org\/images\/core\/emoji\/13.1.0\/72x72\/1f4ac.png\")
Question<\/strong>: How to plot the data so that ![\"\"](\"https:\/\/blog.finxter.com\/wp-content\/uploads\/2022\/04\/image-61.png\")
<\/figure>\n<\/div>\n