Copy-Paste Imputation (CPI) for Energy Time Series

This repository contains the Python implementation of the Copy-Paste Imputation (CPI) method presented in the following paper:

M. Weber, M. Turowski, H. K. Çakmak, R. Mikut, U. Kühnapfel and V. Hagenmeyer, 2021, "Data-Driven Copy-Paste Imputation for Energy Time Series," in IEEE Transactions on Smart Grid, 12, 6, pp. 5409–5419, doi: 10.1109/TSG.2021.3101831.

Installation

To install this project, perform the following steps:

1. Clone the project
2. Open a terminal of the virtual environment where you want to use the project
3. cd into the cloned directory
4. pip install . or pip install -e . to install the project editable.
   • Use pip install -e .[dev] to install with development dependencies

Use

from cpiets.cpi import CopyPasteImputation
import pandas as pd

cpi = CopyPasteImputation()
data = pd.read_csv('data.csv')
cpi.fit(data)
result = cpi.impute()

Input Data Requirements

Example data:

The names of the columns are arbitrary.

Assumptions:

• There are no missing values (nan) at the start or end of the time series.
• A day starts with the first value after 0:00 (0:15 in the example above) and ends with 0:00.
• The time series starts at the beginning of a day and ends at the end of a day.

Supported time formats:

• %Y-%m-%d %H:%M:%S (2020-01-17 13:37:42)
• %d-%b-%Y %H:%M:%S (17-Jan-2020 13:37:42)

Example

In this repository, we included example data derived from the ElectricityLoadDiagrams20112014 data set.

To run the CPI method with simple test data, you can run the example

python example/simple_imputation.py

and play around with the parameters.

Funding

This project is supported by the Helmholtz Association under the Joint Initiative "Energy System 2050 - A Contribution of the Research Field Energy".

License

This code is licensed under the LGPL-3.0 License.