Libraries
First, we need to load a few libraries:
- matplotlib: for creating the chart
- pandas: for data manipulation
import pandas as pd
import matplotlib.pyplot as pltDataset
Our dataset is about the number of people having a certain name, and the evolution of this number over the years. Let's load it with pandas:
# import dataset
df = pd.read_csv("https://raw.githubusercontent.com/holtzy/data_to_viz/master/Example_dataset/5_OneCatSevNumOrdered.csv")
df.head()| year | sex | name | n | prop | |
|---|---|---|---|---|---|
| 0 | 1880 | F | Helen | 636 | 0.006516 |
| 1 | 1880 | F | Amanda | 241 | 0.002469 |
| 2 | 1880 | F | Betty | 117 | 0.001199 |
| 3 | 1880 | F | Dorothy | 112 | 0.001147 |
| 4 | 1880 | F | Linda | 27 | 0.000277 |
Our goal here is to explore the evolution of the number of people named "Ashley" and "Amanda" over the years in a single scatter plot.
For this, we need to change a bit the dataset:
- filter on those 2 names
- filter on only date after 1970
- pivot the table to have the names as columns and the years as rows, using the
pivot_table()function
# filter data
df = df.loc[(df.name=="Ashley") | (df.name=="Amanda")]
df = df.loc[(df.sex=="F") & (df.year>1970)]
df = pd.pivot_table(df, values='n', index=['year'], columns=['name'])
df.head()| name | Amanda | Ashley |
|---|---|---|
| year | ||
| 1971 | 4133.0 | 1164.0 |
| 1972 | 4181.0 | 1176.0 |
| 1973 | 5627.0 | 1253.0 |
| 1974 | 7476.0 | 1626.0 |
| 1975 | 12653.0 | 1988.0 |
Connected scatterplot for evolution
In practice, we just have to call the plot() function with our columns as arguments. We specify that markers have to be used with marker='o' and that the lines have to be connected with linestyle='-'.
fig, ax = plt.subplots(figsize=(8, 6))
# plot the connected scatterplot
ax.plot(df.Amanda, df.Ashley, linestyle='-', marker='o')
# x axis label
plt.xlabel('Amanda')
# y axis label
plt.ylabel('Ashley')
# show the graph
plt.show()
Annotations
Our last chart has a major issue: we don't know which year is represented by each point.
To fix this, we can use the annotate() function to add a text next to each point. However, it's not necessary to add all the years, and decide to only plot 1 out of 3 years.
In practice, we loop over the number of rows with a step of 3 (thanks to range(0, len(df), 3)) and add an annotation with the annotate() function.
fig, ax = plt.subplots(figsize=(8, 6))
# plot the connected scatterplot
ax.plot(df.Amanda, df.Ashley, linestyle='-', marker='o')
# add annotations in every 3 data points with a loop
for line in range(0, df.shape[0], 3):
ax.annotate(
df.index[line],
(df.Amanda.iloc[line], df.Ashley.iloc[line]+1000) ,
va='bottom',
ha='center'
)
# labels and display
plt.xlabel('Amanda')
plt.ylabel('Ashley')
plt.show()
What a nice way to visualize the evolution of the number of people named Ashley and Amanda over the years!
Going further
This post explains how to create a connected scatterplot with matplotlib.
You might be interested in how to reproduce this beautiful connected scatter plot and how to create multiple connected scatter plots on the same chart.
