Code Editor for app.py

import streamlit as st
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Set page title
st.title("CSV Data Visualizer")

# Text field for URL input
url = st.text_input(
    "Enter CSV URL",
    value="https://raw.githubusercontent.com/mwaskom/seaborn-data/master/penguins.csv"
)

# Cache data loading to improve performance
@st.cache_data
def load_data(url):
    try:
        # NOTE: Using corsproxy.io because we're in a WASM environment. If running locally,
        # you can remove the corsproxy.io prefix. Some websites don't work with the proxy,
        # in those cases try removing the proxy prefix.
        if not url.startswith(("http://", "https://")):
            st.error("Please enter a valid URL starting with http:// or https://")
            return None
            
        proxy_url = f"https://corsproxy.io/?{url}" if "corsproxy.io" not in url else url
        data = pd.read_csv(proxy_url)
        return data
    except Exception as e:
        st.error(f"Error loading data: {e}")
        return None

# Load data
data = load_data(url)

if data is not None:
    # Display dataset info
    st.write(f"Dataset shape: {data.shape[0]} rows, {data.shape[1]} columns")
    
    # Control options
    col1, col2 = st.columns(2)
    
    with col1:
        # Number of rows to display
        num_rows = st.slider("Number of rows to display", 5, min(100, data.shape[0]), 10)
    
    with col2:
        # Option to shuffle data
        shuffle = st.checkbox("Randomly shuffle data")
    
    # Process data based on user options
    if shuffle:
        display_data = data.sample(frac=1, random_state=42).reset_index(drop=True)
    else:
        display_data = data.copy()
    
    # Display the dataframe
    st.subheader("Data Preview")
    st.dataframe(display_data.head(num_rows))
    
    # Column selection for visualization
    st.subheader("Data Visualization")
    
    # Get numerical and categorical columns for selection
    numerical_cols = data.select_dtypes(include=['int64', 'float64']).columns.tolist()
    categorical_cols = data.select_dtypes(include=['object', 'category']).columns.tolist()
    all_cols = numerical_cols + categorical_cols
    
    # Let user select columns for visualization
    st.write("Select variables to visualize:")
    col1, col2 = st.columns(2)
    
    with col1:
        x_var = st.selectbox("X-axis variable", all_cols, index=0)
    
    with col2:
        y_var = st.selectbox("Y-axis variable (optional)", ["None"] + all_cols, index=0)
    
    # Determine the best visualization based on selected variables
    if y_var == "None":
        # Single variable visualization
        st.subheader(f"Visualization of {x_var}")
        
        fig, ax = plt.subplots(figsize=(10, 6))
        
        if x_var in numerical_cols:
            # For numerical variables: histogram and KDE
            sns.histplot(data=data, x=x_var, kde=True, ax=ax)
            ax.set_title(f"Distribution of {x_var}")
        else:
            # For categorical variables: count plot
            sns.countplot(data=data, x=x_var, ax=ax)
            ax.set_title(f"Count of {x_var}")
            plt.xticks(rotation=45)
        
        st.pyplot(fig)
        
        # Show summary statistics
        if x_var in numerical_cols:
            st.write("Summary statistics:")
            st.write(data[x_var].describe())
    else:
        # Two variables visualization
        st.subheader(f"Visualization of {x_var} vs {y_var}")
        
        fig, ax = plt.subplots(figsize=(10, 6))
        
        if (x_var in numerical_cols) and (y_var in numerical_cols):
            # Both numerical: scatter plot
            sns.scatterplot(data=data, x=x_var, y=y_var, ax=ax)
            ax.set_title(f"Scatter plot of {x_var} vs {y_var}")
            
            # Add correlation info
            corr = data[[x_var, y_var]].corr().iloc[0, 1]
            st.write(f"Correlation between {x_var} and {y_var}: {corr:.2f}")
            
        elif (x_var in categorical_cols) and (y_var in numerical_cols):
            # Categorical x, Numerical y: box plot
            sns.boxplot(data=data, x=x_var, y=y_var, ax=ax)
            ax.set_title(f"Box plot of {y_var} by {x_var}")
            plt.xticks(rotation=45)
            
        elif (x_var in numerical_cols) and (y_var in categorical_cols):
            # Numerical x, Categorical y: box plot (flipped)
            sns.boxplot(data=data, x=y_var, y=x_var, ax=ax)
            ax.set_title(f"Box plot of {x_var} by {y_var}")
            plt.xticks(rotation=45)
            
        else:
            # Both categorical: count plot
            crosstab = pd.crosstab(data[y_var], data[x_var])
            crosstab.plot(kind='bar', stacked=True, ax=ax)
            ax.set_title(f"Stacked bar plot of {x_var} and {y_var}")
            plt.xticks(rotation=45)
        
        st.pyplot(fig)
else:
    st.warning("Please enter a valid URL to a CSV file.")