Cleaning the rental dataframe and identifying property types

gistfile1.txt

def createPropertyTypeCol(rental_df):
    """This function adds a column called property_type_class to the dataframe 
    Args:
        rental_df ([type]): [The dataframe has a column called property_type]

    Returns:
        [type]: [Dataframe with a more concise list of property types]
    """

    # Property types Private room and Shared Room identified
    property_df = rental_df[["property_type"]].copy()
    property_df.loc[
        property_df["property_type"].str.contains("Private room|Room in"),
        "property_type",
    ] = "Private Room"
    property_df.loc[
        property_df["property_type"].str.contains("Shared room"), "property_type"
    ] = "Shared Room"

    # Extract the second half of all "Entire" property types to get the actual type such as house..
    property_df.loc[
        property_df["property_type"].str.contains("Entire "), "property_type"
    ] = (
        property_df.loc[
            property_df["property_type"].str.contains("Entire "), "property_type"
        ]
        .str.replace("Entire ", "")
        .str.capitalize()
    )

    # We will call Bungalows, Tiny House, Cabins as Cottages:  A house is pretty much any structure where people live, usually one or more families.
    # As per definition:
    #But traditionally, a cabin is a small house built with simple tools
    #A bungalow is a style of house or cottage
    property_df.loc[
        property_df["property_type"].str.contains("Tiny house|Bungalow|Cabin"), "property_type"
    ] = "Cottage"
        
    rental_df["property_type_class"] = property_df["property_type"]

    return rental_df


##############################################################################################################
# Clean up the dataframe
##############################################################################################################
def cleanRentalDF(filename):

    # Read the data into a dataframe
    full_df = pd.read_csv(filename)
    # select out the columns we are interested in
    rental_df = full_df[
        [
            "id",
            "price",
            "listing_url",
            "host_id",
            "host_response_rate",
            "host_response_time",
            "host_acceptance_rate",
            "review_scores_communication",
            "review_scores_location",
            "review_scores_value",
            "review_scores_checkin",
            "reviews_per_month",
            "review_scores_cleanliness",
            "license",
            "instant_bookable",
            "number_of_reviews",
            "first_review",
            "last_review",
            "neighbourhood_cleansed",
            "neighbourhood_group_cleansed",
            "latitude",
            "longitude",
            "accommodates",
            "bathrooms_text",
            "property_type",
            "has_availability",
            "availability_30",
            "availability_60",
            "availability_90",
            "availability_365",
        ]
    ].copy()

    # Make price a float column
    rental_df["price"] = (
        rental_df["price"].str.replace("$", "").str.replace(",", "").astype("float64")
    )

    # Change host response rate from string to float so that it is a continuous value
    # Mean of their reponse rate for the rest
   
    # Convert the response rate to float
    rental_df["host_response_rate_percent"] = (
        rental_df["host_response_rate"].str.replace("%", "").astype("float64")
    )
    rental_df["host_response_rate_percent"] = rental_df.groupby(["host_id"])[
        "host_response_rate_percent"
    ].transform(lambda x: x.fillna(x.mean()))
    # All the values that are still Nan, we do not have any info about and so fill with zero
    rental_df["host_response_rate_percent"] = rental_df[
        "host_response_rate_percent"
    ].fillna(0)
    rental_df = rental_df.drop("host_response_rate", axis="columns")

    # Change response time to one within a dict
    rank_response_time = {
        "within an hour": 1,
        "within a few hours": 2,
        "within a day": 3,
        "a few days or more": 4,
    }
    rental_df["host_reponse_time_rank"] = rental_df["host_response_time"].map(
        rank_response_time
    )
    rental_df["host_reponse_time_rank"] = rental_df["host_reponse_time_rank"].fillna(0)
    rental_df = rental_df.drop("host_response_time", axis="columns")

    # Use the same logic as host_reponse_rate for host_acceptance_rate
    rental_df["host_acceptance_rate_percent"] = (
        rental_df["host_acceptance_rate"].str.replace("%", "").astype("float64")
    )
    rental_df["host_acceptance_rate_percent"] = rental_df.groupby(["host_id"])[
        "host_acceptance_rate_percent"
    ].transform(lambda x: x.fillna(x.mean()))
    # All the values that are still Nan, we do not have any info about and so fill with zero
    rental_df["host_acceptance_rate_percent"] = rental_df[
        "host_acceptance_rate_percent"
    ].fillna(0)
    rental_df = rental_df.drop("host_acceptance_rate", axis="columns")

    # (‘t’ means available and ‘f’ means not available)
    # *Convert t (*true) = 1 , f (false) = 0

    availability_code_dict = {
        "t": 1,
        "f": 0,
    }

    rental_df["instant_bookable"] = rental_df["instant_bookable"].map(
        availability_code_dict
    )
    rental_df["has_availability"] = rental_df["has_availability"].map(
        availability_code_dict
    )

    rental_df["first_review"] = pd.to_datetime(rental_df["first_review"])
    rental_df["last_review"] = pd.to_datetime(rental_df["last_review"])

    # Add a column for a smaller list of property types
    rental_df = createPropertyTypeCol(rental_df)

    return rental_df


################################################################################################################
# Call the cleanup function and setup a global dataframe
################################################################################################################
full_df = cleanRentalDF("data\listings_1.csv")