UBER

import pandas as pd
import numpy as np

# Clean columns & dtypes

df = df.copy()
df.columns = (
    df.columns.str.strip()
              .str.lower()
              .str.replace(' ', '_')
              .str.replace('#', 'num', regex=False)
)

# Parse datetimes
for c in ['start_date', 'end_date']:
    if c in df.columns:
        df[c] = pd.to_datetime(df[c], errors='coerce')

# Keep ZIPCODES as strings
if 'zip_code' in df.columns:
    df['zip_code'] = df['zip_code'].fillna('Unknown')

df['month'] = df['start_date'].dt.to_period('M').dt.to_timestamp()

top_zips = (df['zip_code']
           .value_counts()
            .head(8)
            .index)

trips_by_zip = (df[df['zip_code'].isin(top_zips)]
               .groupby(['month','zip_code'])
               .size()
               .reset_index(name='num_trips'))

top_stations = (df['start_station']
                .value_counts()
                .head(8)
                .index)

trips_by_station = (df[df['start_station'].isin(top_stations)]
                   .groupby(['month', 'start_station'])
                   .size()
                   .reset_index(name='num_trips'))

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

base = df.copy()

clean_map = {c: re.sub(r'\s+', ' ',c).strip() for c in base.columns}
base.rename(columns=clean_map, inplace=True)
base.columns = (base.columns
               .str.lower()
               .str.replace(r'[^0-9a-z]+', '_', regex=True)
                .str.strip('_'))

print("Columns:", base.columns.tolist())

base['start_date'] = pd.to_datetime(base['start_date'], errors='coerce')
subs = base[base['subscriber_type'].astype('string').str.strip().str.lower()==
'subscriber'].copy()

subs['month'] = subs['start_date'].dt.to_period('M').dt.to_timestamp()
trips_by_zip = (subs.groupby(['month','zip_code']).size().reset_index(name='num_trips'))

top_zips = subs['zip_code'].value_counts().head(8).index
trips_by_zip = trips_by_zip[trips_by_zip['zip_code'].isin(top_zips)]

plt.figure(figsize=(12,6))
sns.lineplot(data=trips_by_zip, x='month', y='num_trips', hue='zip_code', marker='o')
plt.title("Subscriber Trips per Month by Zip Code")
plt.xlabel("Month"); plt.ylabel("Number of Trips")
plt.xticks(rotation=45); plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt

top_start = df['start_station'].value_counts().head(10)
plt.figure(figsize=(8,6))
ax = top_start.plot(kind='barh', title='Top 10 Start Stations', color= "steelblue")
for i, value in enumerate(top_start.values):
    ax.text(value + 50, i, str(value), va='center')

plt.xlabel("Number of Trips")
plt.ylabel("Start Station")
plt.tight_layout()
plt.show()
    
top_end = df['end_station'].value_counts().head(10)
plt.figure(figsize=(8,6))
ax = top_end.plot(kind='barh', title='Top 10 End Stations', color= "darkorange")

for i, value in enumerate(top_end.values):
    ax.text(value + 50, i, str(value), va='center')

plt.xlabel("Number of Trips")
plt.ylabel("End Station")
plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt
import pandas as pd


fact_trips = fact_trips.rename(columns=lambda x: x.strip().lower().replace(" ", "_"))

subs = df[df['subscriber_type'] == 'Subscriber']

zip_counts = subs['zip_code'].value_counts()

plt.figure(figsize=(8,6))
ax = zip_counts.head(10).plot(kind='barh', color="steelblue", title="Most Popular Subscriber Zip Codes")
for i, value in enumerate(zip_counts.head(10).values):
    ax.text(value + 5, i, str(value), va='center')
plt.xlabel("Number of Trips")
plt.ylabel("Zip Code")
plt.tight_layout()
plt.show()

plt.figure(figsize=(8,6))
ax = zip_counts.tail(10).plot(kind='barh', color="darkorange", title="Least Common Subscriber Zip Codes")
for i, value in enumerate(zip_counts.tail(10).values):
    ax.text(value + 5, i,str(value), va='center')
plt.xlabel("Number of Trips")
plt.ylabel("Zip Code")
plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt
subs = base[base['subscriber_type'].astype('string').str.strip().str.lower()==
'subscriber'].copy()
zip_counts = subs['zip_code'].value_counts(dropna=True)
low = zip_counts.quantile(0.10)
high = zip_counts.quantile(0.40)

mid_to_bottom = zip_counts[(zip_counts > low) & (zip_counts <= high)]
N = 20
mid_to_bottom_sorted = mid_to_bottom.sort_values(ascending=True).head(N)

plt.figure(figsize=(10,8))
ax = mid_to_bottom_sorted.plot(kind='barh', color='teal')

plt.title("Subscriber Zip Codes (10th-40th Percentile Usage), Sample of 20)", fontsize=14)
plt.xlabel("Number of Trips", fontsize=12)
plt.ylabel("Zip Code", fontsize=12)

for i, v in enumerate(mid_to_bottom_sorted):
    ax.text(v + 0.5, i, str(int(v)), va='center', fontsize=10)
plt.tight_layout()
plt.show()

sns.countplot(data=df, x='subscriber_type')
plt.title("Trips by Subscriber Type vs Customers")
plt.show()
            

sub_by_month = df.groupby(['month', 'subscriber_type']).size().reset_index(name='num_trips')
sns.lineplot(data=sub_by_month, x='month', y='num_trips', hue='subscriber_type')
plt.title("Subscriber vs Customer Trips Over Time")
plt.show()

bike_usage = ( df.groupby(['bike_num', 'start_station', 'zip_code']).size().reset_index(name='num_trips'))
q1 = bike_usage['num_trips'].quantile(0.25)
q2 = bike_usage['num_trips'].quantile(0.50)
q3 = bike_usage['num_trips'].quantile(0.75)

least_used = bike_usage[bike_usage['num_trips'] <= q1]
mid_used = bike_usage[(bike_usage['num_trips'] > q1) & (bike_usage['num_trips'] <= q3)]
most_used = bike_usage[bike_usage['num_trips']> q3]
print("Least used bikes (bottom 25%):")
print(least_used.sort_values('num_trips').head(10))
print("\nMost used bikes (top 25%):")
print(most_used.sort_values('num_trips', ascending=False).head(10))

import matplotlib.pyplot as plt
import seaborn as sns

bike_usage = df.groupby('bike_num').size().reset_index(name='num_trips')
q1 = bike_usage['num_trips'].quantile(0.25)
q2 = bike_usage['num_trips'].quantile(0.50)
q3 = bike_usage['num_trips'].quantile(0.75)

def label_quartile(x):
    if x <= q1: return "Q1(Least Used)"
    elif x <= q2: return "Q2"
    elif x <= q3: return "Q3"
    else: return "Q4(Most Used)"

bike_usage['quartile'] = bike_usage['num_trips'].apply(label_quartile)

quartile_share = bike_usage.groupby('quartile')['num_trips'].sum()/bike_usage['num_trips'].sum()
quartile_order = ["Q1(Least Used)", "Q2", "Q3", "Q4(Most Used)"]


plt.figure(figsize=(10,6))
ax=sns.boxplot(data=bike_usage, x='quartile', y='num_trips',
           order=quartile_order,
           palette="Set2")

for i, quart in enumerate(quartile_order):
    pct=quartile_share.loc[quart]*100
    y_max=bike_usage.loc[bike_usage['quartile']==quart,'num_trips'].max()
    ax.text(i,y_max + 10, f"{pct:.1f}%",
           ha='center', va='bottom', fontsize=12, fontweight='bold')
    
plt.title("Bike Usage Distribution by Quartile", fontsize=14)
plt.xlabel("Quartile"); plt.ylabel("Number of Trips")
plt.tight_layout()
plt.show()

quartile_share = bike_usage.groupby('quartile')['num_trips'].sum() / bike_usage['num_trips'].sum()
print(quartile_share)

import re
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

df0 = df.copy()
df0.columns = (df0.columns
               .str.strip()
               .str.lower()
               .str.replace(r'[^0-9a-z]+', '_', regex=True)
               .str.strip('_'))

# FIX 1: Corrected regex pattern (added missing closing parenthesis)
z = (df0['zip_code'].astype('string')
                     .str.extract(r'(\d{3,5})', expand=False)
                     .str.zfill(5))
df0['zip_clean'] = z.where(z.notna(), np.nan)

subs = df0[df0['subscriber_type'].astype('string').str.strip().str.lower() == 'subscriber'].copy()
subs['area_group'] = np.where(subs['zip_clean'].str.startswith('941', na=False),
                              'SF Core', 'Suburban/Out-of-area')

# FIX 2: Typo in 'reindex'
area_counts = subs['area_group'].value_counts().reindex(['SF Core', 'Suburban/Out-of-area']).fillna(0)

plt.figure(figsize=(6, 5))
bars = plt.bar(area_counts.index, area_counts.values)
plt.title('Trips by Area Group (Subscribers)')
for b in bars:
    y = b.get_height()
    plt.text(b.get_x() + b.get_width() / 2, y + max(area_counts.values) * 0.01,
             f'{int(y):,}', ha='center', va='bottom')
plt.tight_layout()
plt.show()

zip_counts = (subs['zip_clean']
              .value_counts(dropna=True)
              .sort_values(ascending=False))
TOP_N = 25
zip_counts_top = zip_counts.head(TOP_N)
colors = ['C0' if str(z).startswith('941') else 'C1' for z in zip_counts_top.index]

plt.figure(figsize=(10, 8))
# FIX 3: Typo in 'arange'
ypos = np.arange(len(zip_counts_top))[::-1]
plt.barh(zip_counts_top.index[::-1], zip_counts_top.values[::-1], color=colors[::-1])
plt.title(f'Top {TOP_N} Subscriber ZIP Codes\n(SF Core highlighted)')
plt.xlabel('Number of Trips')
plt.ylabel('ZIP Code')

vals = zip_counts_top.values[::-1]
labels = zip_counts_top.index[::-1]
for i, v in enumerate(vals):
    plt.text(v + max(vals) * 0.01, i, f'{int(v):,}', va='center')

from matplotlib.patches import Patch
legend_elems = [Patch(facecolor='C0', label='SF Core (941xx)'),
                Patch(facecolor='C1', label='Suburban/Out-of-area')]
plt.legend(handles=legend_elems, loc='lower right')
plt.tight_layout()
plt.show()

subs['start_date'] = pd.to_datetime(subs['start_date'], errors='coerce')
subs['month'] = subs['start_date'].dt.to_period('M').dt.to_timestamp()

trend = (subs.groupby(['month', 'area_group'])
         .size()
         .reset_index(name='num_trips'))

plt.figure(figsize=(12, 6))
for grp, gdf in trend.groupby('area_group'):
    plt.plot(gdf['month'], gdf['num_trips'], marker='o', label=grp)
plt.title('Monthly Trips by Area Group (Subscribers)')
plt.xlabel('Month')
plt.ylabel('Number of Trips')
plt.xticks(rotation=45)
plt.legend()
plt.tight_layout()
plt.show()

import seaborn as sns
import matplotlib.pyplot as plt

# Bike usage counts
bike_usage = ( df.groupby(['bike_num', 'start_station', 'end_station']).size().reset_index(name='num_trips'))
q3 = bike_usage['num_trips'].quantile(0.75)
top_bikes = bike_usage[bike_usage['num_trips']>= q3]

top_start_stations = (df['start_station'].value_counts().head(10).index)
top_end_stations = (df['end_station'].value_counts().head(10).index)

top_bike_starts = top_bikes[top_bikes['start_station'].isin(top_start_stations)]
top_bike_ends = top_bikes[top_bikes['end_station'].isin(top_end_stations)]

start_summary = (top_bike_starts.groupby('start_station')['num_trips']
                .sum()
                .sort_values(ascending=False)
                )
end_summary = (top_bike_ends.groupby('end_station')['num_trips']
              .sum()
              .sort_values(ascending=False)
              )
fig,axes = plt.subplots(1,2, figsize=(16,6), sharey=True)
sns.barplot(x=start_summary.values, y=start_summary.index, ax=axes[0],color='steelblue')
axes[0].set_title("Top 25% Bikes at Top 10 Start Stations")
axes[0].set_xlabel("Number of Trips"); axes[0].set_ylabel("Start Station")
sns.barplot(x=end_summary.values, y=end_summary.index, ax=axes[1], color='darkorange')
axes[1].set_title("Top 25% Bikes at Top 10 End Stations")
axes[1].set_xlabel("Bumber of Trips");axes[1].set_ylabel("End Station")
plt.tight_layout()
plt.show()
                                                        

duration_summary = df.groupby('Subscriber Type')['Duration'].describe()
print(duration_summary)

import seaborn as sns
import matplotlib.pyplot as plt  # Fixed typo here

plt.figure(figsize=(8,6))
sns.boxplot(data=df, x='subscriber_type', y='duration')
plt.ylim(0,2000)
plt.title("Trip Duration by Subscriber Type")
plt.show()

plt.figure(figsize=(8,6))
sns.violinplot(data=df, x='subscriber_type', y='duration', cut=0, inner="quartile")  # Fixed syntax here
plt.ylim(0,2000)
plt.title("Trip Duration Distribution by Subscriber Type")
plt.show()

from scipy.stats import ttest_ind

subs = df[df['subscriber_type'] == 'Subscriber']['duration']
cust = df[df['subscriber_type'] == 'Customer']['duration']

t_stat, p_val = ttest_ind(subs, cust, equal_var=False)
print(f"T-stat: {t_stat:.2f}, p-value: {p_val:.4f}")