Matplotlib.pyplot散点图设置ylim最小值为0无效问题

Let's break down why your scatter plot is getting distorted and how to fix it:

Problem Overview #

You're pulling AFL player Nat Fyfe's Disposals stats and trying to visualize them with a scatter plot, but setting the y-axis lower limit to 0 (via plt.ylim(0,50) or plt.ylim(ymin=0)) causes the chart to look compressed and warped. Here's your original code for reference:

import pandas as pd
import matplotlib.pyplot as plt
stat_dict={'Disposals' : 0, 'Kicks' : 1, 'Marks' : 2, 'Handballs' : 3, 'Goals' : 4, 'Behinds' : 5, 'Hitouts' : 6, 'Tackles' : 7, 'Rebounds' : 8, 'Inside50s' : 9, 'Clearances': 9, 'Clangers' : 10, 'FreesFor' : 11, 'FreesAgainst' : 12, 'ContestedPosessions' : 13, 'UncontestedPosesseions' : 14, 'ContestedMarks' : 15, 'MarksInside50' : 16, 'OnePercenters' : 17, 'Bounces' : 18, 'GoalAssists' : 19, 'Timeplayed' : 20}
team_lower_case='fremantle'
player="Fyfe, Nat"
stat_required='Disposals'
rounds=7
tables = pd.read_html("https://afltables.com/afl/stats/teams/" +str(team_lower_case)+"/2018_gbg.html")
for df in tables:
    df.drop(df.columns[rounds+1:], axis=1, inplace=True) # remove unwanted columns
    df.columns = df.columns.droplevel(0) # remove extra index level
stat_table=tables[stat_dict[stat_required]]
player_stat=stat_table[stat_table["Player"]==player]
player_stat=player_stat.iloc[:,1:8]
plt.scatter(range(1,rounds+1),player_stat)
plt.ylim(0,50) # or plt.ylim(ymin=0)
plt.show()

Root Cause #

The issue comes down to data dimensions. When you run player_stat=player_stat.iloc[:,1:8], you end up with a 2-dimensional pandas DataFrame (even though it only has one row of data). plt.scatter treats each column of this DataFrame as a separate dataset, so it's actually plotting 7 overlapping scatter series on the same graph. When you lock the y-axis to start at 0, all these overlapping points create the "compressed" visual effect.

Solution #

We just need to convert the 2D DataFrame into a 1D array/Series so plt.scatter only plots one set of points. Here's the fixed code with key improvements:

import pandas as pd
import matplotlib.pyplot as plt

stat_dict={'Disposals' : 0, 'Kicks' : 1, 'Marks' : 2, 'Handballs' : 3, 'Goals' : 4, 'Behinds' : 5, 'Hitouts' : 6, 'Tackles' : 7, 'Rebounds' : 8, 'Inside50s' : 9, 'Clearances': 9, 'Clangers' : 10, 'FreesFor' : 11, 'FreesAgainst' : 12, 'ContestedPosessions' : 13, 'UncontestedPosesseions' : 14, 'ContestedMarks' : 15, 'MarksInside50' : 16, 'OnePercenters' : 17, 'Bounces' : 18, 'GoalAssists' : 19, 'Timeplayed' : 20}
team_lower_case='fremantle'
player="Fyfe, Nat"
stat_required='Disposals'
rounds=7

tables = pd.read_html("https://afltables.com/afl/stats/teams/" +str(team_lower_case)+"/2018_gbg.html")
for df in tables:
    df.drop(df.columns[rounds+1:], axis=1, inplace=True) # Remove extra columns
    df.columns = df.columns.droplevel(0) # Clean up redundant header level

stat_table=tables[stat_dict[stat_required]]
player_stat=stat_table[stat_table["Player"]==player]
# Convert 2D DataFrame to 1D array - this is the critical fix!
player_stat_values = player_stat.iloc[:,1:8].values.flatten()

# Plot the scatter graph with clean 1D data
plt.scatter(range(1, rounds+1), player_stat_values)
# Set y-axis limit without distortion
plt.ylim(0, 50)
# Add labels and title to make the chart easier to interpret
plt.xlabel('Round Number')
plt.ylabel('Disposals')
plt.title(f'{player} 2018 AFL Disposals (Rounds 1-{rounds})')
plt.show()

Key Changes Explained #

1. Data Dimension Fix: Using .values.flatten() converts your single-row DataFrame into a 1D numpy array. This ensures plt.scatter plots only one series of points, not seven overlapping ones.
2. Chart Readability: Added axis labels and a descriptive title to make the graph more intuitive.
3. Proper Y-Axis Limiting: With only one data series plotted, setting plt.ylim(0,50) constrains the axis correctly without causing visual compression.

After these tweaks, your scatter plot will display normally with the y-axis starting at 0, no distortion included.

内容的提问来源于stack exchange，提问作者Kamila Ambro