Code Editor for app.py

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

# Set page title
st.title("Penguin Data Explorer")

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

# Cache data loading function
@st.cache_data
def load_data(url):
    try:
        df = pd.read_csv(url)
        return df
    except Exception as e:
        st.error(f"Error loading data: {e}")
        return None

# Load data
if url:
    data = load_data(url)
    
    if data is not None:
        # Display basic information
        st.write(f"Dataset shape: {data.shape}")
        
        # Add controls in sidebar
        st.sidebar.header("Data Controls")
        
        # Control for shuffling data
        shuffle = st.sidebar.checkbox("Randomly shuffle data", value=False)
        
        # Apply shuffling if selected
        if shuffle:
            data = data.sample(frac=1, random_state=42).reset_index(drop=True)
        
        # Control for number of rows to display
        num_rows = st.sidebar.slider("Number of rows to display", 
                                      min_value=1, 
                                      max_value=len(data), 
                                      value=min(10, len(data)))
        
        # Display the dataframe with selected number of rows
        st.subheader("Data Preview")
        st.dataframe(data.head(num_rows))
        
        # Data visualization section
        st.subheader("Data Visualization")
        
        # Get column names for selection
        numeric_cols = data.select_dtypes(include=['float64', 'int64']).columns.tolist()
        categorical_cols = data.select_dtypes(include=['object', 'category']).columns.tolist()
        all_cols = data.columns.tolist()
        
        # Allow user to select variables for visualization
        st.sidebar.header("Visualization Controls")
        
        # Option to select 1 or 2 variables
        var_count = st.sidebar.radio("Select number of variables to visualize", [1, 2])
        
        if var_count == 1:
            # Single variable visualization
            variable = st.sidebar.selectbox("Select a variable", all_cols)
            
            st.write(f"Visualization for: {variable}")
            
            # Determine best plot based on data type
            if variable in numeric_cols:
                # For numeric data, show histogram and basic stats
                fig, ax = plt.subplots(1, 2, figsize=(12, 5))
                
                # Histogram
                sns.histplot(data[variable].dropna(), kde=True, ax=ax[0])
                ax[0].set_title(f"Distribution of {variable}")
                
                # Boxplot
                sns.boxplot(y=data[variable].dropna(), ax=ax[1])
                ax[1].set_title(f"Boxplot of {variable}")
                
                st.pyplot(fig)
                
                # Display basic statistics
                st.write("Summary statistics:")
                st.write(data[variable].describe())
                
            else:
                # For categorical data, show count plot
                fig, ax = plt.subplots(figsize=(10, 6))
                value_counts = data[variable].value_counts()
                
                sns.countplot(y=data[variable], order=value_counts.index, ax=ax)
                ax.set_title(f"Count of {variable}")
                
                st.pyplot(fig)
                
                # Display value counts
                st.write("Value counts:")
                st.write(value_counts)
        
        else:
            # Two variables visualization
            var1 = st.sidebar.selectbox("Select first variable", all_cols, key="var1")
            var2 = st.sidebar.selectbox("Select second variable", all_cols, key="var2")
            
            st.write(f"Visualization for: {var1} vs {var2}")
            
            # Determine plot type based on variable types
            if (var1 in numeric_cols) and (var2 in numeric_cols):
                # Both numeric - scatter plot with regression line
                fig, ax = plt.subplots(figsize=(10, 6))
                sns.scatterplot(data=data, x=var1, y=var2, ax=ax)
                sns.regplot(data=data, x=var1, y=var2, scatter=False, ax=ax, color='red')
                ax.set_title(f"{var1} vs {var2}")
                
                # Calculate and display correlation
                correlation = data[[var1, var2]].corr().iloc[0, 1]
                st.write(f"Correlation coefficient: {correlation:.4f}")
                
            elif (var1 in categorical_cols) and (var2 in numeric_cols):
                # Categorical vs Numeric - box plot
                fig, ax = plt.subplots(figsize=(10, 6))
                sns.boxplot(data=data, x=var1, y=var2, ax=ax)
                ax.set_title(f"{var2} by {var1}")
                
            elif (var1 in numeric_cols) and (var2 in categorical_cols):
                # Numeric vs Categorical - box plot (swapped)
                fig, ax = plt.subplots(figsize=(10, 6))
                sns.boxplot(data=data, x=var2, y=var1, ax=ax)
                ax.set_title(f"{var1} by {var2}")
                
            else:
                # Both categorical - heatmap of counts
                fig, ax = plt.subplots(figsize=(10, 6))
                
                # Create cross-tabulation
                crosstab = pd.crosstab(data[var1], data[var2], normalize=False)
                sns.heatmap(crosstab, annot=True, fmt='d', cmap='Blues', ax=ax)
                ax.set_title(f"Counts of {var1} vs {var2}")
            
            st.pyplot(fig)
    else:
        st.warning("Please enter a valid CSV URL")