git clone https://github.com/Manuel3567/master-thesis.git
cd master-thesis

python -m venv .venv
.venv\Scripts\activate

pip install -r requirements_gpu.txt

pip install -r requirements.txt

pip install -e .

project_root/
├── data/
│   ├── entsoe/               # Raw aggregated wind power data
│   ├── netztransparenz/      # wind power data of wind park data 
│   └── opendatasoft/         # PLZ list of Germany
|   └── open_meteo/           # Wind speed data

project_root/
├── data/
│   ├── entsoe/               
│   │   ├── Actual Generation per Production Type_201601010000-201701010000.csv
│   │   ├── Actual Generation per Production Type_201701010000-201801010000.csv
│   │   ├── ...
│   │   └── Actual Generation per Production Type_202401010000-202501010000.csv
│   ├── netztransparenz/      
│   │   ├── 50Hertz_Transmission_GmbH_EEG-Zahlungen_Stammdaten_2023.csv
│   │   └── anlagenstammdaten_legende.xslx
│   ├── open_meteo/          
│   │   └── historical/
|   |       └── top_10_biggest_wind_parks_50hertz.json
│   └── opendatasoft/         
│       └── georef-germany-postleitzahl.csv

output_dir = "C:\Users\Manuel\Documents\results"
id=1
from analysis.baseline_model import *
run_baseline_model(id, output_dir)

from analysis.ngboost import *
from analysis.datasets import load_entsoe
output_dir = "C:\Users\Manuel\Documents\results"
entsoe = load_entsoe()

evaluate_ngboost_model(
    entsoe, 
    target_column='power', 
    dist=Normal, 
    case=1, 
    n_estimators=100, 
    learning_rate=0.03, 
    random_state=42, 
    output_file=output_dir,
    train_start = "2016-01-01",
    train_end = "2022-12-31",
    validation_start = "2023-01-01",
    validation_end = "2023-12-31"
)

id=1
output_dir = "C:\Users\Manuel\Documents\results"
from analysis.TabPFN import *
run_tabpfn(id, output_dir)

id=1
output_dir = "C:\Users\Manuel\Documents\results"
from analysis.baseline_model import *
calculate_scores_baseline(id, output_dir)

evaluation is included in evaluate_ngboost_model() 

id=1
output_dir = "C:\Users\Manuel\Documents\results"
from analysis.tabpfn import *
calculate_scores_tabpfn(id, output_dir)

├── output_dir/
│   ├── baseline/
│   │   ├── experiment_1.pkl
│   │   ├── ...
│   │   ├── experiment_2.pkl
│   │   └── quantiles/
│   │       ├── experiment_results_1.pkl
│   │       └── experiment_results_2.pkl
│   │       └── ...
│   ├── ngboost/
│   │   └── full_year/
│   │       ├── case1.xlsx
│   │       ├── case2.xlsx
│   │       ├── ...
│   │       ├── Merged_sheet.xlsx
│   │   └── q4_train/
│   │       ├── case1.xlsx
│   │       ├── case2.xlsx
│   │       ├── ...
│   │       ├── Merged_sheet.xlsx
│   └── tabpfn/
│       ├── experiment_1.pkl
│       ├── experiment_2.pkl
│       ├── ...
│       └── quantiles/
│           ├── experiment_results_1.pkl
│           └── experiment_results_2.pkl
│           └── ...

from analysis.datasets import get_coordinates_of_grid_operators
from analysis.plots import plot_installed_capacity_scatter
aggregated_df = get_coordinates_of_grid_operators()
plot_installed_capacity_scatter(aggregated_df)

from analysis.preprocessor import *
import matplotlib.pyplot as plt
import seaborn as sns

preprocessor = DataPreprocessor()
preprocessor.load_data()
entsoe = preprocessor.df

plt.figure(figsize=(6, 4))
# Plot histogram and KDE
sns.histplot(entsoe['power'], bins=150, color="lightblue")
plt.title("Marginal Distribution of Power")
plt.xlabel("Power [MW]")
plt.ylabel("Frequency")
plt.legend()
plt.show()

from analysis.preprocessor import *
import matplotlib.pyplot as plt
import seaborn as sns

preprocessor = DataPreprocessor()
preprocessor.load_data()
preprocessor = preprocessor.transform_power()
entsoe = preprocessor.df

plt.figure(figsize=(6, 4))
sns.histplot(entsoe['power'], bins=150, color="lightblue")
plt.title("Marginal Distribution of Log Power")
plt.xlabel("ln(Power/Power_max + eps)")
plt.ylabel("Frequency")
plt.legend()
plt.show()