82 lines
2.3 KiB
Python
82 lines
2.3 KiB
Python
import pandas as pd
|
|
import numpy as np
|
|
import matplotlib.pyplot as plt
|
|
|
|
col_names = ['T(C)', 'Td(C)', 'HR%', 'ff(kmh)', 'Gust(kmh)', 'P_0(HPa)', 'P_sea(HPa)']
|
|
|
|
data = pd.read_csv(
|
|
"data_meteo.txt",
|
|
sep=r'\s+',
|
|
skiprows=1,
|
|
usecols=[2, 3, 4, 5, 6, 7, 8],
|
|
names=col_names,
|
|
engine='python'
|
|
)
|
|
|
|
print("загружено записей (строк):", len(data))
|
|
|
|
data = data.apply(pd.to_numeric, errors='coerce')
|
|
n_before = len(data)
|
|
data = data.dropna(subset=col_names)
|
|
n_after = len(data)
|
|
|
|
print(f"после приведения к числам и dropna: {n_after} строк (удалено {n_before - n_after})")
|
|
|
|
def correlation(vec1, vec2, center=False):
|
|
v1 = np.asarray(vec1, dtype=float)
|
|
v2 = np.asarray(vec2, dtype=float)
|
|
|
|
if center:
|
|
v1 = v1 - np.mean(v1)
|
|
v2 = v2 - np.mean(v2)
|
|
|
|
dot_product = np.sum(v1 * v2)
|
|
norm1 = np.sqrt(np.sum(v1 ** 2))
|
|
norm2 = np.sqrt(np.sum(v2 ** 2))
|
|
|
|
if norm1 == 0 or norm2 == 0:
|
|
return np.nan
|
|
|
|
return dot_product / (norm1 * norm2)
|
|
|
|
n = len(col_names)
|
|
|
|
raw_matrix = np.zeros((n, n))
|
|
center_matrix = np.zeros((n, n))
|
|
|
|
for i in range(n):
|
|
for j in range(n):
|
|
r_raw = correlation(data.iloc[:, i], data.iloc[:, j], center=False)
|
|
r_center = correlation(data.iloc[:, i], data.iloc[:, j], center=True)
|
|
|
|
raw_matrix[i, j] = np.round(r_raw, 3) if not np.isnan(r_raw) else np.nan
|
|
center_matrix[i, j] = np.round(r_center, 3) if not np.isnan(r_center) else np.nan
|
|
|
|
df_raw = pd.DataFrame(raw_matrix, index=col_names, columns=col_names)
|
|
df_center = pd.DataFrame(center_matrix, index=col_names, columns=col_names)
|
|
|
|
print("\nкорреляция (raw):")
|
|
print(df_raw.to_string())
|
|
|
|
print("\nкорреляция (centered):")
|
|
print(df_center.to_string())
|
|
|
|
print("\nразница (centered - raw):")
|
|
print((df_center - df_raw).round(3).to_string())
|
|
|
|
plt.figure(figsize=(14, 8))
|
|
|
|
for col in col_names:
|
|
centered_values = data[col] - data[col].mean()
|
|
plt.plot(centered_values.values, label=col, marker='.', linewidth=1, markersize=4)
|
|
|
|
plt.axhline(0, linestyle='--')
|
|
plt.title("центрированные метеоданные")
|
|
plt.xlabel("номер измерения")
|
|
plt.ylabel("отклонение от среднего")
|
|
plt.legend()
|
|
plt.grid()
|
|
plt.tight_layout()
|
|
|
|
plt.savefig("meteo_analysis.png", dpi=150)
|
|
plt.show() |