Introduction:
This Jupyter notebook explores the identification of different day types based on electricity load patterns in an Algerian city. We will analyze a dataset containing hourly recordings of Maximum Power Demand (PMA) and Temperature for two years, from January 1st, 2016, to December 31st, 2017. By applying data mining techniques, we aim to:
- Preprocess and explore the data:
- Analyze descriptive statistics and visualize trends in PMA and Temperature.
- Extract additional features from the date information, including day of week, week of year, month, and holiday identification.
- Identify day types:
- Implement and compare clustering algorithms, such as K-means and Hierarchical clustering, to group days based on their similarity in PMA and Temperature patterns.
- Evaluate the performance of each algorithm using metrics like silhouette score and Calinski-Harabasz index.
- Analyze and interpret the results:
- Characterize the identified day types based on their PMA and Temperature distributions.
- Discuss potential implications and applications of the findings for electricity load forecasting and grid management.
Data Source:
The dataset used in this analysis is stored in a file named pma.xlsx. It contains three columns:
time: Date and time (including hour)pma: Maximum Power Demand (MW)tmp: Temperature (°C)
Software and Tools:
This project will utilize Python libraries such as:
pandasfor data manipulation and analysisnumpyfor scientific computingmatplotlibandseabornfor data visualizationscikit-learnfor machine learning and clustering algorithms