import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv('/Users/danielforsyth/Desktop/bikes.csv')
df.head()

df = df[['LOCATION_B','THEFT_DATE','THEFT_YEAR','THEFT_HOUR','STOLEN_VAL','UCR']]
df.columns = ['Location','Date','Year','Hour','Value','Value_Code']
df.head()

pd.options.display.mpl_style = 'default'
from matplotlib import rcParams
rcParams['figure.figsize'] = (20, 8)
rcParams['figure.dpi'] = 300

df['Date'] = pd.to_datetime(pd.Series(df['Date']))
df.set_index('Date', drop=False, inplace=True)

df['Thefts'] = df['Date'].value_counts()

plt.figure()
plt.ylim(0,20)

df['Thefts'].plot(x_compat=True)


temp = pd.read_csv('/Users/danielforsyth/Desktop/climate.csv',
            parse_dates={'Date': ['MO','DAY','YEAR']},
            usecols=['MO', 'DAY', 'YEAR', 'HIGH'])
temp.head()



temp.plot()

#merge df and temp
merged = pd.merge(df, temp, on='Date', how='outer')
merged.columns = ['Location','Date','Year','Hour','Value','Value_Code','Thefts','Temp']

merged.head()

new = merged.sort('Thefts', ascending=False)
new = new.dropna()
new.head()

thefts = merged['Date'].value_counts()

thefts.head(10)


#thefts per year 2013 , missing 3 months (october-december)
yearly = df.Year.value_counts()
yearly

#Thefts per Month
df['Date'] = df['Date'].map(lambda x: str(x)[:2])
df['Date'] = df['Date'].map(lambda x: x.rstrip('/'))

monthly = df.Date.value_counts()
monthly


df.Hour.value_counts()

df.Value.value_counts()

df.Value_Code.value_counts()

loc = df.Location.value_counts()
loc.head(25)


value = df[['Location','Value_Code']]
value.head()