CODEWITHSUNDEEP
pythonmatplotlibplotgrouping
JK7
JK7

Reputation: 125

change x-axis of a plot

I am trying to create a visualization of vehicles passing by in the first 25 weeks of the years 2015-2020 all in one graph (one curve for every year).

df_data_groups = df_data[(df_data['week']<=25)].groupby(['year','week'])
df_data_weekly = df_data_groups[['NO','nr_of_vehicles']].mean()

fig, ax = plt.subplots()
bp = df_data_weekly['nr_of_vehicles'].groupby('year').plot(ax=ax)

The following is what i get

plot

The x-axis is not right. It should not contain the year, only the weeks, but I don't know how to solve this correctly. It also is not allowing me to create a legend to show which lines belongs to the color of the line, by using:

bp.set_legend()

Upvotes: 0

Views: 150

Answers (1)

JohanC
JohanC

Reputation: 80574

The index shown, is the index of the last dataframe in the group. This dataframe has a 2-level index: the year and the week. Dropping the first index (the year) will only show the week:

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

df_data = pd.DataFrame({'year': np.repeat(np.arange(2015, 2021), 52),
                        'week': np.tile(np.arange(1, 53), 6),
                        'nr_of_vehicles': 200_000 + np.random.randint(-9_000, 10_000, 52 * 6).cumsum()})
df_data_groups = df_data[(df_data['week'] <= 25)].groupby(['year', 'week'])
df_data_weekly = df_data_groups[['nr_of_vehicles']].mean()

fig, ax = plt.subplots()
for year, df in df_data_weekly['nr_of_vehicles'].groupby('year'):
    df.reset_index(level=0, drop=True).plot(ax=ax, label=year)
ax.legend()
ax.margins(x=0.02)
plt.show()

example plot

PS: Note that in the question's code, bp is a list of axes, one ax per year. In this case, all of them point to the same ax. bp is organized as a pandas Series, to obtain the legend, get one of the axes: bp[2015].legend() (or bp.iloc[0].legend()).

Upvotes: 1

Related Questions