import pandas as pd
import numpy
import plotly.plotly as py
from plotly.graph_objs import *

red_orange = '#FF3300'
light_cyan = '#00CCFF'
gray_light = '#bbbbbb'

# takes very long time!
df = pd.read_csv('raw_data/Baltimore_Parking_Citations_2015-03-06.csv',index_col='citation', parse_dates=[10], infer_datetime_format=True)

df['violFine_parsed'] = df['violFine'].str.replace("$","").apply(lambda x: float(x))
df['openFine_parsed'] = df['openFine'].str.replace("$","").apply(lambda x: float(x))
df['openPenalty_parsed'] = df['openPenalty'].str.replace("$","").apply(lambda x: float(x))
df['balance_parsed'] = df['balance'].str.replace("$","").apply(lambda x: float(x))

len(df)

df.head()

groupby_desc = df.groupby('Description')

citation_desc = pd.DataFrame()

# Total fines for each citation type
citation_desc['fines'] = groupby_desc['violFine_parsed'].sum()

# Number of unique locations each citation type was issued at
citation_desc['unique_locs'] = groupby_desc['location'].nunique()

# Fine type concentration by location
citation_desc['fine_conc_loc'] = (citation_desc['fines'] / citation_desc['unique_locs']).apply(lambda x: round(x,2))

citation_desc.sort('fines',inplace=True,ascending=False)

citation_desc['first_date'] = ''
citation_desc['last_date'] = ''
for row in citation_desc.index:
    # First issue date for each citation type
    citation_desc['first_date'].ix[row] = min(groupby_desc['violDate'].get_group(row))
    # Last issue date for each citation type
    citation_desc['last_date'].ix[row] = max(groupby_desc['violDate'].get_group(row))
# Total date range in years for each citation type
citation_desc['date_range_years'] = (citation_desc['last_date'] - citation_desc['first_date']).astype('timedelta64[D]')/365. 
# Fines per year for each citation type
citation_desc['fines_annual'] = citation_desc['fines']/citation_desc['date_range_years']

citation_desc.head()
#citation_desc.to_csv('citation_desc_date_range_new.csv')

# takes a long time!
df_time = df.fillna(0).set_index(df['violDate'])
# takes a long time!
df_time['ImportDate_parsed'] = pd.to_datetime(df_time['ImportDate'])
df_time.sort(inplace=True)
df_time.head()

# Monthly totals for all citations
df_time['violFine'].resample('M',how='count').plot()

# Plot top 10 infraction types over time

fig = plt.figure(figsize=[10,8])
for key in citation_desc[:10].index:
    #print key
    df_time['violFine_parsed'][df_time['Description']==key].resample('M', how='count').plot(label=key,grid='off')

xlabel('')
ylabel('Number of Citations')
legend(loc='upper left')

#fig.savefig('Top 10 Monthly Citation Plot 2015-01-21.png', bbox_tight='inches')

# Monthly mobile speed cameras
df_time['violCode'][df_time['Description']=='Mobile Speed Camera'].resample('M',how='count').plot()

# Total value of fines issued in database before January 21, 2013
df_time['violFine_parsed'][df_time.index <= pd.to_datetime('21-jan-2013')].sum()

df_time[(df_time.index <= pd.to_datetime('21-jan-2013')) & (df_time['balance_parsed']==0)]

df_time[(df_time['ImportDate_parsed'] <= pd.to_datetime('21-jan-2013')) & (df_time['balance_parsed']==0)]

df_time_rec = df_time.ix['21-jan-2013':'21-jan-2015']

total_balance = df_time['balance_parsed'].sum()
total_fines = df_time['violFine_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()
total_pen = df_time['openPenalty_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()



print total_balance, total_fines, total_pen

total_fines+total_pen

(total_fines+total_pen)/total_balance

total_pen/total_balance

df_time['balance_parsed'][df_time['balance_parsed']==0].count()

df_time['openFine_parsed'].sum()+df_time['openPenalty_parsed'].sum()

df_time['balance_parsed'].sum()

df_time['violFine_parsed'].sum()

df_time['openPenalty_parsed'].sum()

df_time['balance_parsed'].sum() - (df_time['openFine_parsed'].sum()+df_time['openPenalty_parsed'].sum())

df_time['violFine_parsed'][df_time['balance_parsed']==0].sum()

df_time['balance_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()

df_time['violFine_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()

df_time['openPenalty_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()

df_time['openFine_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()

df_time['openPenalty_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()+df_time['openFine_parsed'].ix['21-jan-2013':'21-jan-2015'].sum()

df_time['balance_parsed'].ix[:'20-jan-2013'].sum()

df_time['violFine_parsed'].ix[:'20-jan-2013'].sum()

df_time['balance_parsed'].ix[:'20-jan-2013'].sum()/df_time['violFine_parsed'].ix[:'20-jan-2013'].sum()

df_time_2 = df_time.ix['01-feb-2013':'28-feb-2015']
df_time_2['violFine_parsed'].resample('M',how='count').plot()

groupby_desc2 = df_time_2.groupby('Description')

citation_desc2 = pd.DataFrame()

# Total fines for each citation type
citation_desc2['fines'] = groupby_desc2['violFine_parsed'].sum()

# Number of unique locations each citation type was issued at
citation_desc2['unique_locs'] = groupby_desc2['location'].nunique()

# Fine type concentration by location
citation_desc2['fine_conc_loc'] = (citation_desc2['fines'] / citation_desc2['unique_locs']).apply(lambda x: round(x,2))

citation_desc2.sort('fines',inplace=True,ascending=False)

citation_desc2['first_date'] = ''
citation_desc2['last_date'] = ''
for row in citation_desc2.index:
    # First issue date for each citation type
    citation_desc2['first_date'].ix[row] = min(groupby_desc2['violDate'].get_group(row))
    # Last issue date for each citation type
    citation_desc2['last_date'].ix[row] = max(groupby_desc2['violDate'].get_group(row))
# Total date range in years for each citation type
citation_desc2['date_range_years'] = (citation_desc2['last_date'] - citation_desc2['first_date']).astype('timedelta64[D]')/365. 
# Fines per year for each citation type
citation_desc2['fines_annual'] = citation_desc2['fines']/citation_desc2['date_range_years']

citation_desc2
#citation_desc.to_csv('citation_desc_date_range_new.csv')

# Plot top 10 infraction types over time

fig = plt.figure(figsize=[10,8])
ax1 = plt.subplot(111)

for key in citation_desc2[:10].index:
    #print key
    df_time_2['violFine_parsed'][df_time_2['Description']==key].ix['01-feb-2013':'31-dec-2014'].resample('M', how='count').plot(linewidth=2,label=key,grid='off')

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14) 


xlabel('')
ylabel('Number of Citations', fontsize=14)
legend(loc='upper left', fontsize=12)
title('Top 10 Citation Types (by Total Fines)\n 01-Feb-2013 to 31-Dec-2014', fontsize=18)

#fig.savefig('output/Top 10 Monthly Citation Plot 2013-02-01 to 2014-12-31.png', bbox_tight='inches')

groupby_loc = df_time_2.groupby(df_time_2['location'])

loc_counts = groupby_loc['Description'].count()
loc_revenue = groupby_loc['violFine_parsed'].sum()

loc_counts.sort(inplace=True, ascending=False)
loc_revenue.sort(inplace=True, ascending=False)

print '\nNumber of locations:', loc_revenue.count()
print 'Number of locations with fines issued greater than $10,000: ', loc_revenue[loc_revenue>=10000].count()
print '% of all locations with fines issued greater than $10,000:', round(100*float(loc_revenue[loc_revenue>=10000].count()) / float(loc_revenue.count()),1)
print '% Revenue generated by locations with fines issued greater than $10,000:', round(100*loc_revenue[loc_revenue>=10000].sum() / loc_revenue.sum(),1),'\n'


loc_revenue.hist(bins=50)

loc_revenue[:20].to_csv('output/Top20_Target_locations.csv')

df_time_2[:-30]



fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(111)
plot(df_time_2.resample('M', how='count').index,
     df_time_2['tag'][df_time_2['openFine_parsed']==0].resample('M', how='count')/df_time_2['tag'].resample('M', how='count'),
     linewidth=2, color=red_orange)

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

plt.ylim([0,1])
plt.xticks(rotation=45)
# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

title('% Monthly Citations Records With No Open Balance\n', fontsize=16)
plt.subplots_adjust(top=0.9, bottom=0.15)

fig.savefig('output/Monthly_Records_Paid.png', bbox_tight='inches')

red_orange = '#FF3300'

fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(111)
plot(df_time_2.resample('M', how='count').index,
     df_time_2['balance_parsed'][df_time_2['violFine_parsed']>0].resample('M', how='sum')/df_time_2['violFine_parsed'].resample('M', how='sum'),
     linewidth=2, color=red_orange)

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

plt.ylim([0,1.1])
plt.xticks(rotation=45)
# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

title('% Monthly Citations Records With No Open Balance\n', fontsize=18)

df_time_2.resample('M', how='count').index

block_1200_Charles = groupby_loc.get_group('1200 CHARLES ST').groupby('Description')

print 'Fines Issued at 1200 N. Charles St.\n'

block_1200_Charles_fines = pd.DataFrame(columns=['total_fines_issued'])

for key, group in block_1200_Charles:
    
    #print key, block_1200_Charles['violFine_parsed'].get_group(key).sum()

    block_1200_Charles_fines.ix[key] = block_1200_Charles['violFine_parsed'].get_group(key).sum()
    
block_1200_Charles_fines.sort('total_fines_issued',ascending=False, inplace=True)
block_1200_Charles_fines.to_csv('output/1200_Block_N_Charles.csv')    
block_1200_Charles_fines

block_1200_Charles['violFine_parsed'].get_group(key).sum()





groupby_loc.get_group('1200 CHARLES ST').groupby('Description').count()

(df_time_2['violFine_parsed'].resample('D',how='sum')/df_time_2['balance_parsed'].resample('D',how='sum')).plot()

(df_time_2['balance_parsed'].resample('M',how='sum')/df_time_2['violFine_parsed'].resample('M',how='sum')).plot()

(df_time_2['balance_parsed'][df_time_2['balance_parsed']==0].resample('M',how='count')/df_time_2['violFine_parsed'].resample('M',how='count')).plot()

df_time_2['balance_parsed'][df_time_2['balance_parsed']==0].resample('M',how='count')

df_time_2['violFine_parsed'].resample('M',how='count')

fig = plt.figure()

(df_time_2['openFine_parsed'].resample('D',how='sum')/df_time_2['violFine_parsed'].resample('D',how='sum')).plot()
(df_time_2['openPenalty_parsed'].resample('D',how='sum')/df_time_2['violFine_parsed'].resample('D',how='sum')).plot()

plt.grid('off')

legend(['Total Open Fines / Total Fines','Total Open Penalties / Total Fines'])

fig = plt.figure()

(df_time_2['balance_parsed'][df_time_2['balance_parsed']==0].resample('M',how='count')/df_time_2['violFine_parsed'].resample('M',how='count')).plot()

plt.grid('off')

#legend(['Total Open Fines / Total Fines','Total Open Penalties / Total Fines'])

import plotly.plotly as py

fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(1,1,1)

plot1 = (df_time_2['openFine_parsed'].resample('D',how='sum')/df_time_2['violFine_parsed'].resample('D',how='sum')).plot(color=red_orange)
#(df_time_2['openPenalty_parsed'].resample('D',how='sum')/df_time_2['violFine_parsed'].resample('D',how='sum')).plot()
linear_ext = pd.DataFrame([.03, .11],index=[pd.to_datetime('01-feb-2013'),pd.to_datetime('20-dec-2013')])
plot2 = linear_ext[0].plot(color=red_orange, linestyle='--', linewidth=3)


# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

plt.ylim([0,1.1])
#plt.xticks(rotation=45)
# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

plt.title('% Daily Total Fines Issued Remaining Unpaid', fontsize=18)
plt.xlabel('')
plt.grid('off')

#fig.savefig('output/Daily_Total_Fines_Unpaid.png', bbox_tight='inches')
#py.iplot_mpl(fig)

fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(1,1,1)

plot1 = (df_time_2['openFine_parsed'][df_time_2['openFine_parsed']>0].resample('D',how='count')/df_time_2['violFine_parsed'].resample('D',how='count')).plot(color=red_orange)
#(df_time_2['openPenalty_parsed'].resample('D',how='sum')/df_time_2['violFine_parsed'].resample('D',how='sum')).plot()
linear_ext = pd.DataFrame([.03, .11],index=[pd.to_datetime('01-feb-2013'),pd.to_datetime('20-dec-2013')])
plot2 = linear_ext[0].plot(color=red_orange, linestyle='--', linewidth=3)


# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

plt.ylim([0,1.1])
#plt.xticks(rotation=45)
# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

plt.title('% Daily Number Citations Issued Remaining Unpaid', fontsize=18)
plt.xlabel('')
plt.grid('off')

#fig.savefig('output/Daily_Citations_Unpaid.png', bbox_tight='inches')

df_time_2['openFine_parsed'].resample('D',how='count')
#df_time_2['openFine_parsed'][df_time_2['openFine_parsed']>0].resample('D',how='count')

df_time_2['city'] = 'Baltimore'
df_time_2['state'] = 'MD'
#df_time_2[['location','city','state']].ix['2013'].to_csv('output/2013_citation_locations.csv')
#df_time_2[['location','city','state']].ix['2014'].to_csv('output/2014_citation_locations.csv')

# Random samples for mapping purposes
N = 1000

df_2013 = df_time_2[['tag','location','city','state']].ix['2013']
df_2014 = df_time_2[['tag','location','city','state']].ix['2014']

sample_2013 = df_2013.iloc[np.random.choice(len(df_2013), N, replace=False)]
#sample_2013.to_csv('output/2013_random_sample.csv')

sample_2014 = df_2014.iloc[np.random.choice(len(df_2014), N, replace=False)]
#sample_2014.to_csv('output/2014_random_sample.csv')


block_1200_Charles['violFine_parsed'].plot()

df_time_2['violFine_parsed'][df_time_2['location']=='1200 CHARLES ST'].resample('M',how='sum').plot()



np.std(df_time_2['violFine_parsed'].ix['2014'].resample('D',how='sum'))

np.std(df_time_2['violFine_parsed'].ix['2013'].resample('D',how='sum'))

# Overall monthly citation increase
fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(1,1,1)

plot1 = df_time_2['violFine_parsed'].ix[:'dec-2014'].resample('M',how='count').plot(color=red_orange, grid='off',linewidth=3)

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

#plt.ylim([0,1.1])
#plt.xticks(rotation=45)
# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

plt.ylabel('Total Monthly Number of Citations', fontsize=16)
plt.xlabel('')

#fig.savefig('output/Total_Monthly_Citations.png', bbox_tight='inches')

# % Increase Sept 2013 to Mar 2014

print 'Overall Ratio, March 2014 citations to September 2013 citations: ',round(float(df_time_2['violFine_parsed'].ix['mar-2014'].count()) / float(df_time_2['violFine_parsed'].ix['sept-2013'].count()), 2)

# FIRST NEED TO SLICE OUT ONLY LOCATIONS WITH TICKETS IN Sep 2013 and Mar 2014
#sep13_uniq_locs = df_time_2['location'].ix['sep-2013'].unique()
mar14_uniq_locs = df_time_2['location'].ix['mar-2014'].unique()

a = df_time_2[df_time_2['location'].isin(mar14_uniq_locs)]
#b = a[a['location'].isin(mar14_uniq_locs)]

Mar14_Sep13 = pd.DataFrame()
Mar14_Sep13['Mar_2014'] = a['location'].ix['Mar-2014'].value_counts()
Mar14_Sep13['Sept_2013'] = a['location'].ix['Sept-2013'].value_counts()
Mar14_Sep13['Sept_2013'].fillna(0, inplace=True)
Mar14_Sep13['Diff'] = Mar14_Sep13['Mar_2014'] - Mar14_Sep13['Sept_2013']
Mar14_Sep13['Ratio'] = Mar14_Sep13['Mar_2014'] / Mar14_Sep13['Sept_2013']

Mar14_Sep13.sort('Diff', ascending=False, inplace=True)
Mar14_Sep13[:25]
#Mar14_Sep13[:25].to_csv('output/Top_25_Increase_Citations_Sept13_Mar14_Locs.csv')



df_time_2['tag'][df_time_2['location']=='400 LOT JFA'].ix[:'dec-2014'].resample('M', how='count').plot()

a = df_time_2[df_time_2['location']=='400 LOT JFA'].groupby('Description')
a['tag'].count()

df_time_2['tag'][df_time_2['location']=='3200 LOT WAVERLY'].ix[:'dec-2014'].resample('M', how='count').plot()

fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(1,1,1)
plot1 = df_time_2['tag'][df_time_2['location']=='5100 ROLAND AVE'].ix[:'dec-2014'].resample('M', how='count').plot(grid='off', color=red_orange, linewidth=3)

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

#plt.ylim([0,1.1])
#plt.xticks(rotation=45)
# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

plt.ylabel('Total Monthly Number of Citations', fontsize=16)
plt.xlabel('')
plt.title('5100 Roland Ave.', fontsize=16)

#fig.savefig('output/5100_Roland_Citations.png', bbox_tight='inches')

a = df_time_2[df_time_2['location']=='5100 ROLAND AVE'].groupby('Description')
a['tag'].count()

fig = plt.figure(figsize=[10,6])
ax1 = plt.subplot(1,1,1)

# convert resampled index to right date format
my_x_labels = df_time_2['violFine_parsed'][df_time_2['location']=='4200 WICKFORD RD'].resample('M',how='count').index.map(lambda x: x.strftime('%b %Y'))

plot1= bar(xrange(len(my_x_labels)),
            df_time_2['violFine_parsed'][df_time_2['location']=='4200 WICKFORD RD'].resample('M',how='count'),
            align='center',
            color=red_orange)
xlim([0,len(my_x_labels)])

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

plt.xticks(xrange(len(df_time_2['violFine_parsed'][df_time_2['location']=='4200 WICKFORD RD'].resample('M',how='count'))),my_x_labels,ha='center',rotation=90)

# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

#title('4200 Wickford Rd.', fontsize=20)
text(len(my_x_labels)-1,24,'4200 Wickford Rd.',ha='right',fontsize=20)
plt.subplots_adjust(top=0.95, bottom=0.2)
ylabel('Monthly Ticket Count', fontsize=16)
#fig.savefig('output/4200_Wickford.png', bbox_tight='inches')

df_time_2['violFine_parsed'][df_time_2['location']=='4200 WICKFORD RD'].count()#.to_csv('output/4200_Wickford_Citations.csv')

df_time_2['Description'][df_time_2['location']=='4200 WICKFORD RD']#.to_csv('output/4200_Wickford_Citations.csv')

Charles_St = ['900 CHARLES ST',
              '1000 CHARLES ST',
              '1100 CHARLES ST',
              '1200 CHARLES ST',
              '1300 CHARLES ST']

# All days

Charles_St_citations = df_time_2[df_time_2['location'].isin(Charles_St)]

Charles_St_hour_count = Charles_St_citations['tag'].groupby(Charles_St_citations.index.hour).count()
bar(Charles_St_hour_count.index,Charles_St_hour_count)

# Charles St Weekend

Charles_St_citations_weekend = Charles_St_citations.ix[(Charles_St_citations.index.dayofweek==5) | (Charles_St_citations.index.dayofweek==6)]
Charles_St_citations_weekend_hour_count = Charles_St_citations_weekend['tag'].groupby(Charles_St_citations_weekend.index.hour).count()
bar(Charles_St_citations_weekend_hour_count.index,Charles_St_citations_weekend_hour_count,align='center')
xlim([-1,24])
plt.show()
#Charles_St_citations_weekday = Charles_St_citations.index.dayofweek.isin([0,1,2,3,4])

# Charles St Weekday

Charles_St_citations_weekday = Charles_St_citations.ix[Charles_St_citations.index.dayofweek<5]
Charles_St_citations_weekday_hour_count = Charles_St_citations_weekday['tag'].groupby(Charles_St_citations_weekday.index.hour).count()
bar(Charles_St_citations_weekday_hour_count.index,Charles_St_citations_weekday_hour_count,align='center')
xlim([-1,24])
plt.show()

# Total citation counts on 5 blocks of N. Charles St. by day of week
fig = plt.figure(figsize=[8,6])
ax1 = plt.subplot(1,1,1)

my_x_labels = ['Mon','Tues','Weds','Thurs','Fri','Sat','Sun']

a=Charles_St_citations.groupby(Charles_St_citations.index.dayofweek)
a['tag'].count().plot(kind='bar', grid='off', color=red_orange)

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

text(6.25,2500,'N. Charles St. \n Citations',ha='right',va='center',fontsize=20)
ylabel('Citations', fontsize=16)
xticks(range(0,7),my_x_labels, rotation=0)
plt.show()

fig.savefig('output/N_Charles_by_Day.png', bbox_tight='inches')

Charles_St_citations.groupby('Description')['tag'].count()

fig = plt.figure(figsize=[10,8])
ax1 = plt.subplot(1,1,1)

a=pd.DataFrame(pd.Series(Charles_St_citations_weekday.index.hour).value_counts(),index=range(0,23))
b=pd.DataFrame(pd.Series(Charles_St_citations_weekend.index.hour).value_counts(),index=range(0,23))
b.fillna(0,inplace=True)
a.sort(inplace=True)
b.sort(inplace=True)
bar(b.index,b[0],color=light_cyan,align='center')
bar(a.index,a[0],color=gray_light,bottom=b[0],align='center')

xlim([-1,24])
xlabel('Hour of Day', fontsize=16)
ylabel('Citations', fontsize=16)
text(0,2000,'N. Charles St. Citations',ha='left',va='center',fontsize=20)
legend(['Weekend','Weekday'],loc='upper left', fontsize=16, frameon=False,
       bbox_to_anchor=(.14, -.13, 1, 1),
       bbox_transform=gcf().transFigure)

# turn off square border around plot
ax1.spines["top"].set_visible(False)  
ax1.spines["bottom"].set_visible(False)  
ax1.spines["right"].set_visible(False)  
ax1.spines["left"].set_visible(False)

# turn off ticks
ax1.tick_params(axis="both", which="both", bottom="off", top="off",
               labelbottom="on", left="off", right="off", labelleft="on",labelsize=14)

fig.savefig('output/N_Charles_by_Hour.png', bbox_inches='tight')


b

Charles_St_citations