macroecon-tools 1.0.3

Suite of tools for macroeconomics analysis

Macroecon-Tools

A open-source set of tools to assist with macroeconomic work. Extrends the pandas series class into an adjusted timeseries class along with other functionality.

Timeseries and TimeseriesTable Package

This Python package provides two custom classes, Timeseries and TimeseriesTable, for working with time series data and collections of time series. It builds on pandas and numpy to offer additional metadata and advanced operations tailored for macroeconomic analysis of time series

---

Features

Timeseries Class

The Timeseries class extends pandas.Series and adds the following functionality:

• Metadata Tracking: Includes attributes like name, source_freq, data_source, and transformations. Automatically appends the applied functions to transformations.

• Transformation Methods:

  • logdiff: Compute the log difference of the series.
  • diff: Compute the difference over a specified lag.
  • log and log100: Apply logarithmic transformations. (Note: log100 multiplies the log by 100).

• Aggregation:

  • Aggregate data into different timeframes (e.g., monthly, quarterly) using methods like sum, mean, and lastvalue.

• Filtering:

  • Apply linear or Hamilton filtering methods.

• Truncation:

  • Restrict the series to a specific time range.

• Persistence:

  • Save and load timeseries objects using pickle.

• Custom Representation:

  • Enhanced string representations for easier inspection.

TimeseriesTable Class

The TimeseriesTable class manages collections of Timeseries objects:

• Maintains both a dictionary-like structure and a pandas.DataFrame representation.
• Supports operations like truncation, dropping missing values, and correlation computation.
• Facilitates saving and loading the entire TimeseriesTable.

-to_latex:

• Convert the table to a LaTeX table for easy inclusion in reports. -trunc:
• Truncate the table to a specific time range. -corr:
• Compute the correlation matrix of the table. -dropna:
• Drop missing values from the table.

---

Usage

Timeseries Class

import pandas as pd
from timeseries import Timeseries

# Create a sample Timeseries
data = pd.Series([1 + 0.3 * i for i in range(10)], index=pd.date_range('2024-01-01', periods=10))
ts = Timeseries(data, name="Sample Data", source_freq="daily", data_source="Generated")

# Apply transformations
log_diff = ts.logdiff(1)
truncated = ts.trunc('2024-01-01', '2024-01-04')

TimeseriesTable Class

import pandas as pd
from timeseries import TimeseriesTable, Timeseries

# Create a sample TimeseriesTable
table_data = {
    "series1": Timeseries(pd.Series([1.1, 1.2, 1.3], index=pd.date_range('2024-01-01', periods=3))),
    "series2": Timeseries(pd.Series([2.1, 2.2, 2.3], index=pd.date_range('2024-01-01', periods=3))),
}
table = TimeseriesTable(table_data)

# Perform operations
correlation_matrix = table.corr()
truncated_table = table.trunc('2024-01-01', '2024-01-02')