Scenario

A CSV file contains daily temperature readings over multiple years. You need to decompose this time-series data to analyze its underlying patterns.

Task

Write a Python script at /home/interview/decompose_timeseries.py using statsmodels that reads /home/interview/temperature_data.csv, decomposes the temperature time-series into trend, seasonal, and residual components using additive decomposition, and saves each component to separate CSV files: /home/interview/trend.csv, /home/interview/seasonal.csv, and /home/interview/residual.csv.

Note: The statsmodels module is already installed.

Example

Input (temperature_data.csv):

date,temperature
2023-01-01,8.5
2023-01-02,9.2
...

Expected outputs:

• trend.csv - Long-term trend component
• seasonal.csv - Repeating seasonal pattern
• residual.csv - Random fluctuations