Time Series Forecasting with TimeGPT

%%capture
%pip install nixtla>=0.5.1
%pip install yfinance

import pandas as pd
import yfinance as yf
from nixtla import NixtlaClient

import os

timegpt_api_key = os.environ["TIMEGPT_API_KEY"]

# Setup NixtlaClient
nixtla_client = NixtlaClient(api_key = timegpt_api_key)


# Downloading Amazon stock price data
ticker = 'AMZN'
amazon_stock_data = yf.download(ticker)
amazon_stock_data = amazon_stock_data.reset_index()

# Displaying the dataset
amazon_stock_data.head()

amazon_stock_data.describe()

nixtla_client.plot(amazon_stock_data, time_col='Date', target_col='Close')

model = nixtla_client.forecast(df=amazon_stock_data, model='timegpt-1', h=24,freq="B", time_col='Date', target_col='Close')
model.tail()

nixtla_client.plot(amazon_stock_data, model, time_col='Date', target_col='Close',max_insample_length=60 )

import pandas as pd

def read_tsf_from_file(file_path):
    data = []
    start_date = pd.to_datetime('2002-01-01 00:00:00')

    # Open and read the file from the directory
    with open(file_path, 'r') as file:
        for line in file:
            if line.startswith('T'):
                parts = line.strip().split(':')
                unique_id = parts[0] + '-' + parts[1]
                values = list(map(float, parts[3].split(',')[:-1]))
                
                # Generate datetime index at half-hour intervals
                periods = len(values)
                date_range = pd.date_range(start=start_date, periods=periods, freq='30min')
                
                # Append to data list
                for dt, value in zip(date_range, values):
                    data.append([unique_id, dt, value])

    # Convert the list of data into a DataFrame
    return pd.DataFrame(data, columns=['unique_id', 'ds', 'y'])

# Example usage
file_path = 'australian_electricity_demand_dataset.tsf'
demand_df = read_tsf_from_file(file_path)

# Display the dataframe
demand_df.head()

nixtla_client.plot(
    demand_df, 
    max_insample_length=365, 
)

# Detect anomalies
anomalies_df = nixtla_client.detect_anomalies(
    demand_df, 
    time_col='ds', 
    target_col='y', 
    freq='H', 
)

anomalies_df[anomalies_df["anomaly"]==1].head()