AI Data Analyst for Instant Data Insights

import pandas as pd
import numpy as np
import json

# Load data
df = pd.read_csv(DATASET_PATHS[0])

# Parse Date, remove timezone info
# Convert to UTC then drop tz
df['Date'] = pd.to_datetime(df['Date'], utc=True).dt.tz_convert('UTC').dt.tz_localize(None)
df.sort_values('Date', inplace=True)

# Daily returns
df['Return'] = df['Close'].pct_change()
daily_ret = df['Return'].dropna()

# Compute metrics
start_price = df['Close'].iloc[0]
end_price = df['Close'].iloc[-1]

# Duration in years
duration_years = (df['Date'].iloc[-1] - df['Date'].iloc[0]).days / 365.25

cagr = (end_price / start_price) ** (1 / duration_years) - 1
cumulative_return = end_price / start_price - 1
daily_vol = daily_ret.std()
annual_vol = daily_vol * np.sqrt(252)

# Max drawdown
dumax = df['Close'].cummax()
drawdown = df['Close'] / dumax - 1
max_dd = drawdown.min()
max_dd_date = df.loc[drawdown.idxmin(), 'Date'].strftime('%Y-%m-%d')

# 1-year return
one_year_ago = df['Date'].iloc[-1] - pd.Timedelta(days=365)
df_1y = df[df['Date'] >= one_year_ago]
return_1y = df_1y['Close'].iloc[-1] / df_1y['Close'].iloc[0] - 1 if not df_1y.empty else None

# Return since 2016
df_2016 = df[df['Date'] >= pd.Timestamp('2016-01-01')]
return_since_2016 = df_2016['Close'].iloc[-1] / df_2016['Close'].iloc[0] - 1 if not df_2016.empty else None

metrics = {
    'CAGR': cagr,
    'Annualized Volatility': annual_vol,
    'Cumulative Return': cumulative_return,
    'Max Drawdown': max_dd,
    'Max Drawdown Date': max_dd_date,
    '1Y Return': return_1y,
    'Return Since 2016': return_since_2016
}

# Output JSON
print(json.dumps(metrics, indent=2))

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

# Load and clean data
df = pd.read_csv(DATASET_PATHS[0], encoding='ISO-8859-1')
df = df[df['Quantity'] > 0].copy()
df['InvoiceDate'] = pd.to_datetime(df['InvoiceDate'], errors='coerce')
df = df.dropna(subset=['InvoiceDate', 'CustomerID'])

# Cohort analysis
df['CustomerID'] = df['CustomerID'].astype(int)
df['InvoiceMonth'] = df['InvoiceDate'].dt.to_period('M')
first_purchase = df.groupby('CustomerID')['InvoiceMonth'].min().reset_index()
first_purchase.columns = ['CustomerID', 'CohortMonth']

df = df.merge(first_purchase, on='CustomerID')
df['CohortIndex'] = ((df['InvoiceMonth'].dt.year - df['CohortMonth'].dt.year) * 12 + 
                     (df['InvoiceMonth'].dt.month - df['CohortMonth'].dt.month) + 1)

# Calculate retention rates
cohort_data = df.groupby(['CohortMonth', 'CohortIndex'])['CustomerID'].nunique().reset_index()
cohort_counts = cohort_data.pivot(index='CohortMonth', columns='CohortIndex', values='CustomerID')
cohort_sizes = cohort_counts.iloc[:, 0]
retention = cohort_counts.divide(cohort_sizes, axis=0)

# Visualize
plt.figure(figsize=(12, 8))
sns.heatmap(retention, annot=True, fmt='.0%', cmap='Blues')
plt.title('Cohort Retention Rates')
plt.tight_layout()
plt.savefig('cohort_retention_heatmap.png')