nyc_restaurants_1.py

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
pd.set_option("precision", 2)
pd.options.display.float_format = '{:20,.2f}'.format

df = pd.read_csv("Restaurant_Grades.csv")

# Replace spaces with underscores
df.columns = df.columns.str.replace("CUISINE DESCRIPTION","CUISINE")
df.columns = df.columns.str.replace(" ","_")

# Clean up some of the names
df["DBA"].replace("'","", inplace=True) # Remove apostrophe
df["DBA"].replace(" ?\(.+\)", "", regex=True, inplace=True) # Remove values in parenthesis
df["DBA"].replace(" ?#.*", "", regex=True, inplace=True) # Remove # followed by some string

# Convert dates to datetime
df.GRADE_DATE = pd.to_datetime(df.GRADE_DATE, format="%m/%d/%Y")
df.RECORD_DATE = pd.to_datetime(df.RECORD_DATE, format="%m/%d/%Y")

# Create a unique key based on restaurant
df["KEY"] = df[['DBA', 'BUILDING', "STREET", "ZIPCODE"]].astype(str).apply(lambda x: ' '.join(x), axis=1)
print("num ratings: {} num unique restaurants: {}".format(len(df), len(df.KEY.unique())))
# num ratings: 186185 num unique restaurants: 24607

# Scores are broken down by critical, not critical and not applicable, which results in duplicate rows for the same rating
df = df[["KEY","DBA","BORO","CUISINE","SCORE","GRADE","GRADE_DATE","RECORD_DATE"]].drop_duplicates()
df = df.sort_values(["KEY","GRADE_DATE"],ascending=[True,False])