sh16ma

'''
seaborn でのヒートマップ描画
'''
# Correlation Matrix
fig, ax = plt.subplots(figsize=(30, 25))

mat = dt.corr('pearson')
mask = np.triu(np.ones_like(mat, dtype=bool))
cmap = sns.diverging_palette(230, 20, as_cmap=True)

sh16ma

'''
seabornでの描画：ストライプ図、バイオリン図、箱ひげ図
'''
import matplotlib.pyplot as plt
import seaborn as sns

def plots(x, y):
    '''
     x 特徴量、 y 目的変数
    '''

sh16ma

# Looking at NaN % within the data

nan = pd.DataFrame(all_df.isna().sum(), columns=['NaN_sum'])
nan['feat'] = nan.index
nan['ratio(%)'] = (nan['NaN_sum']/all_df.shape[0])*100
nan = nan[nan['NaN_sum'] > 0]
nan = nan.sort_values(by = ['NaN_sum']) #option: ascending=False 昇順
nan['Usability'] = np.where(nan['ratio(%)'] > 20, 'Discard', 'Keep')
nan

sh16ma

# visualization
import matplotlib.pyplot as plt
import seaborn as sns; sns.set()
import missingno as msg

# データフレームを綺麗に出力する関数
import IPython
def display(*dfs, head=True):
    for df in dfs:
        IPython.display.display(df.head() if head else df)

sh16ma

def nan_rank(df,  usabilty=20):
    """
    df : データフレーム
    usabilty : データの割合に応じての足切りライン
    """
    nan = df.isnull().sum().reset_index()
    nan.columns = ["name", "count"]
    nan["ratio"] = (nan["count"] / df.shape[0])*100
    nan["usabilty"] = np.where(nan["ratio"] > usabilty, "Discard", "Keep")
    nan = nan[nan["count"] > 0].sort_values(by="ratio")

sh16ma

from sklearn.metrics import accuracy_score # 正解率
from sklearn.metrics import precision_score # 適合率
from sklearn.metrics import recall_score # 再現率
from sklearn.metrics import f1_score # F値

def evaluations(y_test, y_predict, average) -> dict: # 辞書型で返す
    """
    y_test: 
      予測したいデータ
    y_predict:

sh16ma

%matplotlib inline
import pandas as pd
import matplotlib.pyplot as plt
# preprocessing
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.pipeline import Pipeline
# algorithm
from sklearn.svm import SVC # サポートベクトルマシン
from sklearn.linear_model import LogisticRegression # ロジスティック回帰

sh16ma

# libraries
import numpy as np
from sklearn.metrics import r2_score # 決定係数(R^2)
from sklearn.metrics import mean_squared_error # 平均二乗誤差(MSE) 
from sklearn.metrics import mean_absolute_error # 平気絶対値誤差(MAE)


# ------------------------------
# 　途中式割愛 
# ------------------------------

sh16ma

/*
 * 
 *
 */
function() {
  var read_timer_flg = false;

  // scroll ratio.
  function getScrollAmount() {
      var scroll = window.pageYOffset;

sh16ma

#standaardSQL

select
/**********************************************
 * Google BigQueryでの要約統計量
 **********************************************/
    count(passenger_count) as n
    , avg(passenger_count) as mean
    , stddev(passenger_count) as std
    , min(passenger_count) as min