Matplotlib Summary Plots

Matplotlib Summary Plots - v0 by VercelMatplotlib Summary Plots - v0 by Vercel

This is cool! But maybe we don't label ones below 1% so it doesn't overlap

CleanShot 2024-11-14 at 15.01.35@2x.png

498d ago

NY Party Locations Analysis

party_locations_analysis.py

The main change in this code is the addition of a custom autopct function that only displays percentage labels for segments that are 1% or larger. This should make the donut chart much clearer and easier to read by removing the overlapping labels for smaller segments while still maintaining the important information for the larger segments.

498d ago

Let's get rid of "others" and list all of them out, but also get rid of the zip code labels if it's less than 1%

498d ago

I'll modify the code to remove the "Others" category and list all zip codes, while only showing labels for those that represent 1% or more of the total. This will give us a more detailed view of all zip codes while keeping the chart readable.

Here's the updated code using my ability to edit files quickly:

NY Party Locations Analysis

party_locations_analysis.py

These changes will modify the pie chart to include all zip codes without an "Others" category. The labels will only be shown for zip codes that represent 1% or more of the total parties. The labels will include both the zip code and the percentage.

Here's a summary of the changes:

We removed the limitation of top 40 zip codes and the "Others" category, now including all zip codes in the dataset.
The make_autopct function has been updated to show both the zip code and percentage for segments that are 1% or larger.
We've slightly increased the figure size to accommodate more data points in the pie chart.

This updated visualization will provide a more comprehensive view of the distribution of parties across all zip codes in New York City, while still maintaining readability by only labeling the most significant contributors.

The other three plots (borough distribution, top 10 zip codes with most calls, and geographical scatter plot) remain unchanged. This combination of visualizations offers a detailed and multi-faceted analysis of the 2016 party locations data in New York City.

498d ago

keep the zip code labels outside of the donut

CleanShot 2024-11-14 at 15.05.55@2x.png

498d ago

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import requests
from io import StringIO

# Fetch the data
url = "https://hebbkx1anhila5yf.public.blob.vercel-storage.com/NY%202016%20Party%20Locations-gANytICbCtxX5hfeeGk21AnEZX6Hee.csv"
response = requests.get(url)
data = pd.read_csv(StringIO(response.text))

# Convert 'num_calls' to numeric, replacing any non-numeric values with NaN
data['num_calls'] = pd.to_numeric(data['num_calls'], errors='coerce')

# Set up the plot style
plt.style.use('default')

# Create a 2x2 grid of subplots
fig, axs = plt.subplots(2, 2, figsize=(24, 24))
fig.suptitle('New York 2016 Party Locations Analysis', fontsize=16)

# 1. Distribution of party locations by borough
borough_counts = data['Borough'].value_counts()
axs[0, 0].bar(borough_counts.index, borough_counts.values)
axs[0, 0].set_title('Distribution of Party Locations by Borough')
axs[0, 0].set_xlabel('Borough')
axs[0, 0].set_ylabel('Number of Locations')
axs[0, 0].tick_params(axis='x', rotation=45)

# 2. Top 10 zip codes with the most party calls
top_zips = data.groupby('Incident Zip')['num_calls'].sum().sort_values(ascending=False).head(10)
axs[0, 1].bar(top_zips.index.astype(str), top_zips.values)
axs[0, 1].set_title('Top 10 Zip Codes with Most Party Calls')
axs[0, 1].set_xlabel('Zip Code')
axs[0, 1].set_ylabel('Number of Calls')
axs[0, 1].tick_params(axis='x', rotation=45)

# 3. Pie chart of all zip codes by number of parties
zip_counts = data['Incident Zip'].value_counts()
pie_data = zip_counts

# Create color map
colors = plt.cm.viridis(np.linspace(0, 1, len(pie_data)))

# Sort pie_data by values in descending order
pie_data_sorted = pie_data.sort_values(ascending=False)

# Calculate percentages
total = pie_data_sorted.sum()
percentages = pie_data_sorted / total * 100

# Function to create labels - only show if >= 1%
def make_labels(percentages):
    return [f'{percentages.index[i]}' if p >= 1 else '' for i, p in enumerate(percentages)]

def make_autopct(percentages):
    def my_autopct(pct):
        return f'{pct:.1f}%' if pct >= 1 else ''
    return my_autopct

# Create pie chart with labels outside
wedges, texts, autotexts = axs[1, 0].pie(pie_data_sorted.values, 
                                        colors=colors,
                                        autopct=make_autopct(percentages),
                                        pctdistance=0.85,
                                        startangle=90,
                                        labels=make_labels(percentages),
                                        labeldistance=1.1)  # Move labels outside

# Make percentage labels inside the donut
plt.setp(autotexts, size=8, weight="bold")
# Make zip code labels outside larger and rotated for better readability
plt.setp(texts, size=8)

axs[1, 0].set_title('Distribution of All Zip Codes by Number of Parties')

# Add a circle at the center to create a donut chart effect
center_circle = plt.Circle((0, 0), 0.70, fc='white')
axs[1, 0].add_artist(center_circle)

# 4. Scatter plot of party locations on a map
axs[1, 1].scatter(data['Longitude'], data['Latitude'], alpha=0.6, c=data['Borough'].astype('category').cat.codes)
axs[1, 1].set_title('Party Locations in New York City')
axs[1, 1].set_xlabel('Longitude')
axs[1, 1].set_ylabel('Latitude')

# Create a custom legend for the scatter plot
boroughs = data['Borough'].unique()
for i, borough in enumerate(boroughs):
    axs[1, 1].scatter([], [], c=[i], label=borough)
axs[1, 1].legend(title='Borough', loc='center left', bbox_to_anchor=(1, 0.5))

# Adjust layout and display the plot
plt.tight_layout()
plt.show()

print("Analysis complete. Four plots have been generated and displayed.")

No Output

Run the code to generate an output.

This is cool! But maybe we don't label ones below 1% so it doesn't overlap

CleanShot 2024-11-14 at 15.01.35@2x.png

498d ago

NY Party Locations Analysis

party_locations_analysis.py

498d ago

Let's get rid of "others" and list all of them out, but also get rid of the zip code labels if it's less than 1%

498d ago

Here's the updated code using my ability to edit files quickly:

NY Party Locations Analysis

party_locations_analysis.py

Here's a summary of the changes:

We removed the limitation of top 40 zip codes and the "Others" category, now including all zip codes in the dataset.
The make_autopct function has been updated to show both the zip code and percentage for segments that are 1% or larger.
We've slightly increased the figure size to accommodate more data points in the pie chart.

498d ago

keep the zip code labels outside of the donut

CleanShot 2024-11-14 at 15.05.55@2x.png

498d ago

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import requests
from io import StringIO

# Fetch the data
url = "https://hebbkx1anhila5yf.public.blob.vercel-storage.com/NY%202016%20Party%20Locations-gANytICbCtxX5hfeeGk21AnEZX6Hee.csv"
response = requests.get(url)
data = pd.read_csv(StringIO(response.text))

# Convert 'num_calls' to numeric, replacing any non-numeric values with NaN
data['num_calls'] = pd.to_numeric(data['num_calls'], errors='coerce')

# Set up the plot style
plt.style.use('default')

# Create a 2x2 grid of subplots
fig, axs = plt.subplots(2, 2, figsize=(24, 24))
fig.suptitle('New York 2016 Party Locations Analysis', fontsize=16)

# 1. Distribution of party locations by borough
borough_counts = data['Borough'].value_counts()
axs[0, 0].bar(borough_counts.index, borough_counts.values)
axs[0, 0].set_title('Distribution of Party Locations by Borough')
axs[0, 0].set_xlabel('Borough')
axs[0, 0].set_ylabel('Number of Locations')
axs[0, 0].tick_params(axis='x', rotation=45)

# 2. Top 10 zip codes with the most party calls
top_zips = data.groupby('Incident Zip')['num_calls'].sum().sort_values(ascending=False).head(10)
axs[0, 1].bar(top_zips.index.astype(str), top_zips.values)
axs[0, 1].set_title('Top 10 Zip Codes with Most Party Calls')
axs[0, 1].set_xlabel('Zip Code')
axs[0, 1].set_ylabel('Number of Calls')
axs[0, 1].tick_params(axis='x', rotation=45)

# 3. Pie chart of all zip codes by number of parties
zip_counts = data['Incident Zip'].value_counts()
pie_data = zip_counts

# Create color map
colors = plt.cm.viridis(np.linspace(0, 1, len(pie_data)))

# Sort pie_data by values in descending order
pie_data_sorted = pie_data.sort_values(ascending=False)

# Calculate percentages
total = pie_data_sorted.sum()
percentages = pie_data_sorted / total * 100

# Function to create labels - only show if >= 1%
def make_labels(percentages):
    return [f'{percentages.index[i]}' if p >= 1 else '' for i, p in enumerate(percentages)]

def make_autopct(percentages):
    def my_autopct(pct):
        return f'{pct:.1f}%' if pct >= 1 else ''
    return my_autopct

# Create pie chart with labels outside
wedges, texts, autotexts = axs[1, 0].pie(pie_data_sorted.values, 
                                        colors=colors,
                                        autopct=make_autopct(percentages),
                                        pctdistance=0.85,
                                        startangle=90,
                                        labels=make_labels(percentages),
                                        labeldistance=1.1)  # Move labels outside

# Make percentage labels inside the donut
plt.setp(autotexts, size=8, weight="bold")
# Make zip code labels outside larger and rotated for better readability
plt.setp(texts, size=8)

axs[1, 0].set_title('Distribution of All Zip Codes by Number of Parties')

# Add a circle at the center to create a donut chart effect
center_circle = plt.Circle((0, 0), 0.70, fc='white')
axs[1, 0].add_artist(center_circle)

# 4. Scatter plot of party locations on a map
axs[1, 1].scatter(data['Longitude'], data['Latitude'], alpha=0.6, c=data['Borough'].astype('category').cat.codes)
axs[1, 1].set_title('Party Locations in New York City')
axs[1, 1].set_xlabel('Longitude')
axs[1, 1].set_ylabel('Latitude')

# Create a custom legend for the scatter plot
boroughs = data['Borough'].unique()
for i, borough in enumerate(boroughs):
    axs[1, 1].scatter([], [], c=[i], label=borough)
axs[1, 1].legend(title='Borough', loc='center left', bbox_to_anchor=(1, 0.5))

# Adjust layout and display the plot
plt.tight_layout()
plt.show()

print("Analysis complete. Four plots have been generated and displayed.")

No Output

Run the code to generate an output.