#!/usr/bin/env python
# coding: utf-8

# # 2016 Phillies Games Broadcast on National Television
# 
# I like watching the Phillies.  I do not have cable.  Some Phillies games are broadcast on national television.  This is how I made a list of those games.
# 
# ## [Pandas](http://pandas.pydata.org/)
# 
# Pandas is a data analysis tool for the [Python](https://www.python.org/) programming language.  It can do a tremendous amount of really powerful data analysis and visualization.  It's a gun in this CSV knife fight.

# In[1]:


import pandas as pd


# A downloadable [CSV schedule](http://philadelphia.phillies.mlb.com/schedule/downloadable.jsp#csv-format) is available from [mlb.com](http://mlb.com).  Here is a [direct link](http://mlb.mlb.com/ticketing-client/csv/EventTicketPromotionPrice.tiksrv?team_id=108&display_in=singlegame&ticket_category=Tickets&site_section=Default&sub_category=Default&leave_empty_games=true&event_type=T&event_type=Y) to the Phillies schedule.
# 
# The CSV schedule will be used to instantiate a Pandas [DataFrame](http://pandas.pydata.org/pandas-docs/version/0.13.1/generated/pandas.DataFrame.html) object.

# In[2]:


schedule = pd.DataFrame.from_csv("phillies-2016.csv")


# ## What does the schedule metadata look like?

# In[3]:


schedule.info()


# 190 games and 16 columns of data for each game.  
# 
# ## What does the schedule data itself look like?

# In[4]:


schedule.head()


# ## Cleaning up the schedule
# 
# The `DESCRIPTION` column contains the broadcast information.  Less interesting columns can be removed.

# In[6]:


schedule.drop(["REMINDER OFF", 
             "REMINDER ON",
             "START TIME ET",
             "END DATE",
             "END DATE ET",
             "END TIME",
             "END TIME ET",
             "REMINDER TIME",
             "REMINDER TIME ET",
             "SHOWTIMEAS FREE",
             "SHOWTIMEAS BUSY",
             "REMINDER DATE"], axis=1, inplace=True)
schedule.head()


# ## What are all of the stations that games are broadcast on this season?
# 
# The `DESCRIPTION` column is nice because it mentions the stations that games are broadcast on.  Sometimes a game is broadcast on two channels at once.  There is also radio broadcast information that I'm not interested in right now.

# In[11]:


schedule.DESCRIPTION.head(50)


# ### Parse television station broadcast channels from `DESCRIPTION`
# 
# Thankfully, the `DESCRIPTION` column data is parseable.  Getting a list of television broadcast stations for each game is not _too_ difficult.

# In[73]:


description = schedule.DESCRIPTION[6]
print description


# Grab the rough station string with a regular expression.

# In[123]:


import re

TV_STATION_RE = re.compile(r"""Local\s+TV:\s+    # TV token
                               (?P<stations>.*)  # Group everything following it lazily as stations
                               """, re.X)


# Use that to pull them out and do some text wrangling.

# In[ ]:


def tv_stations_from_description(description):
    """Return a list of television stations embedded in the given description."""
    tv_stations = []
    result = re.search(TV_STATION_RE, str(description))
    if result:
        media_delimiter = "-----"
        tv_station_str = result.group("stations").split(media_delimiter)[0]
        tv_stations = tv_station_str.split("- ")
        tv_stations = [s.strip() for s in tv_stations]
    return tv_stations


# Test it out on all of the descriptions.

# In[126]:


tv_stations = set()
for d in schedule.DESCRIPTION:
    tv_stations |= set(tv_stations_from_description(d))
tv_stations    


# Applying this function to the DataFrame yields a [`Series`](http://pandas.pydata.org/pandas-docs/stable/dsintro.html#series) of all television stations on which the Phillies are broadcast this season.

# In[127]:


stations_series = schedule.DESCRIPTION.apply(lambda d: tv_stations_from_description(d))
stations_series


# Double check the `set` of stations from that `Series`.

# In[129]:


set([station for stations in stations_series.values for station in stations])


# The 190 Phillies games are broadcast on 6 television channels.  Unfortunately only 1 of those 6 stations are available without a cable television subscription.  This means that I can only watch games on NBC.
# 
# ## The Phillies national television broadcast schedule
# 
# Filtering the `DESCRIPTION` column to national television broadcast stations yields only the games which I can watch over the air with my [HD antenna](http://amzn.to/1r5eZmQ).

# In[117]:


national_broadcast_schedule = schedule[schedule.DESCRIPTION.str.contains("NBC 10") == True]
national_broadcast_schedule


# In[118]:


national_broadcast_schedule.describe()


# This means that I have the possibility to watch 10 out of 190 Phillies games this season which is roughly 5%.