平均利用時間行列

gistfile1.txt

# Aligning the station names exactly with the transition probability matrix, excluding any extra stations
# First, we adjust the list of station names to exclude the extra station (if any)
aligned_stations = set(transition_probability_matrix.columns) - {'start_station_name'}

# Filter the station_stats dataframe to include only those rows where both the start and end stations are in the aligned_stations set
filtered_station_stats = station_stats[
    (station_stats['start_station_name'].isin(aligned_stations)) & 
    (station_stats['end_station_name'].isin(aligned_stations))
]

# Creating a pivot table for the mean travel times, exactly aligned with the stations in the transition probability matrix
mean_travel_time_matrix = filtered_station_stats.pivot(
    index='start_station_name', 
    columns='end_station_name', 
    values='mean_travel_time_min'
)

# Filling NaN values with 0, assuming no travel time for non-existent trips
mean_travel_time_matrix = mean_travel_time_matrix.fillna(0)

# Reindexing the matrix to include only the stations from the transition probability matrix, filling missing values with 0
mean_travel_time_matrix = mean_travel_time_matrix.reindex(
    index=aligned_stations, 
    columns=aligned_stations, 
    fill_value=0
)

mean_travel_time_matrix.to_csv(path + 'mean_travel_time_matrix', index=True)

mean_travel_time_matrix