app.py

# Import necessary libraries
import streamlit as st
import pandas as pd
import altair as alt
import matplotlib.pyplot as plt
import seaborn as sns
from scipy.stats import zscore

# Load data
file_path = 'kiva_loans.csv'
df_kiva_loans = pd.read_csv(file_path)

# Clean data
df_kiva_loans = df_kiva_loans.drop(['use', 'disbursed_time', 'funded_time', 'posted_time', 'tags'], axis=1)
df_kiva_loans.dropna(subset=['partner_id', 'borrower_genders'], inplace=True)

# Calculate Z-scores
z_scores = zscore(df_kiva_loans['funded_amount'])
df_kiva_loans['outlier_funded_amount'] = (z_scores > 3) | (z_scores < -3)
df_kiva_loans_cleaned = df_kiva_loans[~df_kiva_loans['outlier_funded_amount']]

# Streamlit App Title
st.title('BDS24_Weekly_Assignment_Week 2 | Tryfonas Karmiris')

# Display the cleaned data table
st.table(df_kiva_loans_cleaned.head())

# Dropdown and slider for Altair chart
st.subheader('Distribution of Funded Amounts')
# Altair chart: simple distribution of funded amounts
chart = alt.Chart(df_kiva_loans_cleaned).mark_bar().encode(
    alt.X('funded_amount', bin=alt.Bin(maxbins=50)),  # Use funded_amount for distribution
    y='count()',
).properties(
    title='Distribution of Funded Amounts'
)
st.altair_chart(chart, use_container_width=True)

# Dropdown and slider for Matplotlib dual-axis plot
st.subheader('Top Values by Selected Variable')

# Dropdown for plot type
plot_type = st.selectbox("Select Variable to Display", ['country', 'repayment_interval', 'sector'])

# Slider to select the number of top values to display
num_columns = st.slider(
    "Select Number of Columns to Display",
    min_value=5,
    max_value=50,
    value=10,  # default value
    step=1
)

# Select the top values based on the selected variable and number of columns
if plot_type == 'country':
    top_values = df_kiva_loans.groupby('country')['funded_amount'].agg(['sum', 'count']).nlargest(num_columns, 'sum').reset_index()
    x_column = 'country'
    count_column = 'count'
elif plot_type == 'repayment_interval':
    top_values = df_kiva_loans.groupby('repayment_interval')['funded_amount'].agg(['sum', 'count']).nlargest(num_columns, 'sum').reset_index()
    x_column = 'repayment_interval'
    count_column = 'count'
else:  # sector
    top_values = df_kiva_loans.groupby('sector')['funded_amount'].agg(['sum', 'count']).nlargest(num_columns, 'sum').reset_index()
    x_column = 'sector'
    count_column = 'count'

# Create a dual-axis bar plot using Matplotlib
fig, ax1 = plt.subplots(figsize=(12, 9))
plt.xticks(rotation=90)

# Bar plot for funded_amount
color = 'tab:blue'
ax1.set_xlabel(x_column.replace("_", " ").title())
ax1.set_ylabel('Funded Amount', color=color)
ax1.bar(top_values[x_column], top_values['sum'], color=color, alpha=0.6, label='Funded Amount')
ax1.tick_params(axis='y', labelcolor=color)

# Create a second y-axis for count
ax2 = ax1.twinx()
color = 'tab:red'
ax2.set_ylabel('Count', color=color)
ax2.plot(top_values[x_column], top_values[count_column], color=color, marker='o', linestyle='-', linewidth=2, label='Count')
ax2.tick_params(axis='y', labelcolor=color)

# Add titles and labels
plt.title(f'Top {num_columns} by {plot_type.replace("_", " ").title()}')
fig.tight_layout()
st.pyplot(fig)

# Boxplot (or Violin Plot) after the dual-axis plot
st.subheader('Funded Amount vs. Selected Variable')

# Filter the data based on the selected variable and number of top values
if plot_type == 'sector':
    top_values_boxplot = df_kiva_loans.groupby('sector')['funded_amount'].agg('sum').nlargest(num_columns).index
    filtered_df_boxplot = df_kiva_loans_cleaned[df_kiva_loans_cleaned['sector'].isin(top_values_boxplot)]
elif plot_type == 'country':
    top_values_boxplot = df_kiva_loans.groupby('country')['funded_amount'].agg('sum').nlargest(num_columns).index
    filtered_df_boxplot = df_kiva_loans_cleaned[df_kiva_loans_cleaned['country'].isin(top_values_boxplot)]
else:  # repayment_interval
    filtered_df_boxplot = df_kiva_loans_cleaned

# Create a boxplot
fig, ax = plt.subplots(figsize=(12, 6))
if plot_type != 'repayment_interval':
    # Use sorted values for 'sector' and 'country'
    top_values_sorted = df_kiva_loans.groupby(plot_type)['funded_amount'].agg('sum').nlargest(num_columns).index
    sns.boxplot(x=plot_type, y='funded_amount', data=filtered_df_boxplot, order=top_values_sorted, ax=ax)
else:
    # No specific sorting needed for 'repayment_interval'
    sns.boxplot(x=plot_type, y='funded_amount', data=filtered_df_boxplot, ax=ax)

plt.title('Funded Amount by Selected Variable')
plt.xlabel(plot_type)
plt.ylabel('Funded Amount')
plt.xticks(rotation=45)
st.pyplot(fig)

# Dropdown for Seaborn countplot
st.subheader('Repayment Interval by Selected Variable')

# Dropdown for selecting variable for Seaborn countplot
plot_var = st.selectbox("Select Variable for Countplot", ['sector', 'country'])

# Slider to select the number of top values to display for Seaborn countplot
num_top_values = st.slider(
    "Select Number of Top Values to Display",
    min_value=5,
    max_value=50,
    value=10,  # default value
    step=1
)

# Filter the data based on the selected variable and number of top values
if plot_var == 'sector':
    top_values_plot = df_kiva_loans.groupby('sector')['funded_amount'].agg('count').nlargest(num_top_values).index
    filtered_df_plot = df_kiva_loans_cleaned[df_kiva_loans_cleaned['sector'].isin(top_values_plot)]
elif plot_var == 'country':
    top_values_plot = df_kiva_loans.groupby('country')['funded_amount'].agg('count').nlargest(num_top_values).index
    filtered_df_plot = df_kiva_loans_cleaned[df_kiva_loans_cleaned['country'].isin(top_values_plot)]

# Create Seaborn countplot
fig, ax = plt.subplots(figsize=(10, 6))
sns.countplot(x='repayment_interval', hue=plot_var, data=filtered_df_plot, ax=ax)
plt.title(f'Repayment Interval by {plot_var.replace("_", " ").title()}')
plt.xlabel('Repayment Interval')
plt.xticks(rotation=90)
plt.ylabel('Count')
plt.legend(title=plot_var.replace("_", " ").title(), bbox_to_anchor=(1.05, 1), loc='upper left')
st.pyplot(fig)