Data analysis parts collection

Data analysis parts collection

--History

• 2021/01/21 ver1

version.cmd

python --version
:: Python 3.7.6
jupyter --version
:: jupyter core     : 4.6.1
:: jupyter-notebook : 6.0.3
:: qtconsole        : 4.6.0
:: ipython          : 7.12.0
:: ipykernel        : 5.1.4
:: jupyter client   : 5.3.4
:: jupyter lab      : 1.2.6
:: nbconvert        : 5.6.1
:: ipywidgets       : 7.5.1
:: nbformat         : 5.0.4
:: traitlets        : 4.3.3

package

import.py

import pandas as pd #A library that provides functions to support data analysis
import numpy as np #Numerical calculation extension module
import matplotlib #Package for data visualization
import matplotlib.pyplot as plt #Interface for automatic plotting, seems
from datetime import datetime as dt #Module for manipulating dates and times
from sklearn.preprocessing import StandardScaler #Module for data standardization

Handle Japanese in matplotlib

Reference source: Japanese with matplotlib

matplot_japanese.py

from matplotlib import rcParams
rcParams['font.family'] = 'sans-serif'
rcParams['font.sans-serif'] = ['Hiragino Maru Gothic Pro', 'Yu Gothic', 'Meirio', 'Takao', 'IPAexGothic', 'IPAPGothic', 'VL PGothic', 'Noto Sans CJK JP']

Size of graph or image to display

figsize.py

plt.figure(figsize=(20,2))

Data input / output

Read from SQL

#SQL
import pymysql
import sqlalchemy
from sqlalchemy import create_engine

#Connection information
url = 'mysql+pymysql://root:[email protected]:3306/databasename?charset=utf8'
engine = sqlalchemy.create_engine(url, echo=False)
#Run
query = "SELECT * FROM Table"
dataset = pd.read_sql(query,con = engine)

CSV input / output

csv.py

dataset = pd.read_csv("pass of csv", encoding = "utf-8")
dataset.to_csv("pass of csv", encoding="shift-jis")

Output to clipboard

clipboard.py

!pip install pyperclip
import pyperclip
pyperclip.copy(STR_XXX)

Pre-processing

Confirmation of contents

info.py

dataset.info()
dataset.describe() #max,min,mean,std,Quartile, etc.

Plastic surgery

index.py

#Indexing data frames
dataset2 = dataset.set_index('StoreCD')
#Type change
dataset['colum']=dataset['colum'].astype(int)

Data frame decomposition and combination

--Decompose the crosstabbed data into simple ROW data

melt_concat.py

#Data frame decomposition
meltDF1 = pd.melt(dataset,id_vars='index_column',var_name='horizon_axis_column_name',value_name='value_column_name')

#Combine data frames
concatDF = pd.concat([meltDF1,meltDF2]) 

Handling null values

null.py

dataset.fillna(0,inplace=True) #0 fill
dataset.isnull() #Check throughout the data frame
dataset.isnull().any() #Check by column

Dummy variable

dummy.py

target_col = 'a'
str_colmns = ['b','c','d'] #Non-numeric column
dummie_cols = ['b'] #Columns that you want to dummy in non-numeric columns
exclude_cols = [col for col in str_colmns if col not in dummie_cols] #
#Dummy variable
df = pd.get_dummies(data=df, columns=dummie_cols)
#Specify the column to be used as a feature from the column after dummyization
feature_cols = [col for col in df.columns if col not in exclude_cols]
　

Visualization and exploration

Take the correlation coefficient between items

corr.py

corrDF = df[feature_cols].corr()

Heat map

heatmap.py

import seaborn as sns
sns.heatmap(corrDF,annot=False) #annot=Display numbers with True

Overlapping bar chart

distplot.py

import seaborn as sns
g=sns.FacetGrid(df,hue="target_column",height=3)
g.map(sns.distplot,"feature_column",kde=False)
g.add_legend()

seaborn color settings

--You can set linked colors

color.py

flatui = ['#969696', '#DA5019']
sns.set_palette(flatui)

comment ――Because it is troublesome to find out from the past code and copy and paste