Use the YAMNet CSV

At the end of the previous example, we downloaded a CSV file with basic information about the YAMNet classifications. In this example, we will demonstrate how to use this file to display part of the data.

NOTE: To reproduce the results in this example, download the YAMNet classification timeline (.csv) in the Additional Products section in this report.

Running the Example

First, we will read the CSV file and display the columns’ titles.

import pandas as pd

# Input path to CSV downloaded from report
input_file: str = "path/to/yamnet/data/YAMNet.csv"

# Load CSV
df = pd.read_csv(input_file)
print(f'Available columns: {df.columns.values}')

Now that we’ve confirmed we can read the CSV, we’ll focus on the classification that YAMNet thinks is the most likely. In the file, this classification is named class_0. It is accompanied by the confidence score score_0.

import pandas as pd
from typing import Dict

# Input path to CSV downloaded from report
input_file: str = "path/to/yamnet/data/YAMNet.csv"

# Load CSV
df = pd.read_csv(input_file)
print(f'Available columns: {df.columns.values}')

# Get ID of stations
stations = df["Station ID"].unique()

# for each of the stations, perform these tasks
for station in stations:
    # Get rows with station ID specific information
    df_station = df.loc[df['Station ID'] == station]
    # only look at the most likely classification
    dict_counts = df_station["class_0"].value_counts().to_dict()

    # Get index of max and min yamnet scores
    idx_max_score = df_station["score_0"].idxmax()
    idx_min_score = df_station["score_0"].idxmin()
    # Get names of most and least frequent class names
    most_frequent_class = df_station["class_0"].mode()[0]
    least_frequent_class = min(dict_counts, key=dict_counts.get)

    # Calculate average yamnet score per classification name
    dict_avg_scores: Dict = {}
    for idx, class_name in enumerate(dict_counts.keys()):

        df_class_name = df_station.loc[df_station['class_0'] == class_name]
        dict_avg_scores[f"{class_name}"] = "{0:.3f}".format(df_class_name['score_0'].mean())

Our statistics have been gathered, so we will now plot the results as a pie chart:

import pandas as pd
import matplotlib.pyplot as plt
from typing import Dict

# Input path to CSV downloaded from report
input_file: str = "path/to/yamnet/data/YAMNet.csv"

# Load CSV
df = pd.read_csv(input_file)
print(f'Available columns: {df.columns.values}')

# Get ID of stations
stations = df["Station ID"].unique()

for station in stations:
    # Get rows with station ID specific information
    df_station = df.loc[df['Station ID'] == station]
    dict_counts = df_station["class_0"].value_counts().to_dict()

    # Some preparation to extract metrics
    # Get index of max and min yamnet scores
    idx_max_score = df_station["score_0"].idxmax()
    idx_min_score = df_station["score_0"].idxmin()
    # Get names of most and least frequent class names
    most_frequent_class = df_station["class_0"].mode()[0]
    least_frequent_class = min(dict_counts, key=dict_counts.get)

    # Calculate average yamnet score per classification name
    dict_avg_scores: Dict = {}
    for idx, class_name in enumerate(dict_counts.keys()):

        df_class_name = df_station.loc[df_station['class_0'] == class_name]
        dict_avg_scores[f"{class_name}"] = "{0:.3f}".format(df_class_name['score_0'].mean())

    # Start pie chart figure
    labels = [f"{key} ({str(dict_counts.get(key))})" for key in dict_counts.keys()]
    sizes = dict_counts.values()

    fig, ax = plt.subplots()
    plt.suptitle(f"Station {station} - Total YAMNet annotations: {sum(dict_counts.values())}")
    ax.pie(sizes, labels=labels, autopct='%1.1f%%', startangle=90)
    ax.axis('equal')  # Equal aspect ratio ensures that pie is drawn as a circle.

plt.show()

Example Output

For a more complete example on how to load the CSV data, visit Github.