mark-clements

## docker command shortlist

docker cp <containerId>:/file/path/within/container /host/path/target
docker stop $(docker container ls -q)
docker stop containername
docker rm containername
docker system prune -a # delete all containers and images
docker container ps -a
docker network ls
docker volume ls -q

mark-clements

######## VIF correlation
from statsmodels.stats.outliers_influence import variance_inflation_factor
# filter of quantitative columsn retained thus far
cols = ['age', 'anc_1er_achat', 'anc_1ere_vis','anc_creation', 
        'anc_der_achat_j', 'anc_der_vis_j', 'ca_t4',  'del_cpt_1er_achat_j', 
        'mnt_achat_max', 'mnt_achat_min', 'mnt_achat_moy',   'nb_achats_t4', 
        'nb_cat_vis',  'nb_sec_vis', 'nb_vis_tot', 'nb_visit_apres-midi_14_18',
        'nb_visit_matin_6_10', 'nb_visit_midi_10_14', 'nb_visit_soir_18_22', 
        'membercreationdate','firstattendingdate']

mark-clements

#############################################
# OneHotEncode Qualitative Variables
#############################################
# When incorporating nominal variables, you must first consider
   whether they have 'directionality'. Does converting 'red', 'green', 'blue',
   to 1, 2, 3 indicate that 'blue' is greater than 'red'? If the answer is no
   you should not simply convert a column via factorize or to type 'category'.
   You should instead use OneHotEncoder or get_dummies, which creates
   seperate columns

mark-clements

        # simple bar  plot
        # bar charts groups are typically categorical variables.
        #  histograms the groups are typically intervals of another continuous variable. 
        ax = r['percentage'].plot(kind="bar")
        ax.set(xlabel='Total Purchases by Customer', ylabel='% of Best = True')
        plt.gca().yaxis.grid(True, alpha = 0.5)
        plt.title('Percentage of total Best==True group by Number of Purchases')
        plt.show
        
        # Barplot of model performance

mark-clements

# Import sqlalchemy's create_engine() function
from sqlalchemy import create_engine

# Create the database engine
engine = create_engine("sqlite:///data.db")

# View the tables in the database
print(engine.table_names())

# Create the database engine

mark-clements

  library(kableExtra)
  
  # KableExtra, example of setting conditional colour with base R
  mtcars$mpg <- cell_spec(mtcars$mpg, color = ifelse(mtcars$mpg > 20, "red", "blue")) 
  
  ################################################################################
  # print each element of a dt to table
  ################################################################################
  ```{r results='asis'}  

mark-clements

# get factor levels used in model
model_step_5$xlevels

# fitted values of model
model$fitted

# predict against new dataset requires the use of 'newdata' or will revert to previous data
predict(model, newdata = test, type = "response")

mark-clements

#### Some tabs in rmarkdown : {.tabset .tabset-fade} 

##### Tab 1
This section must be one level below the tabset header above, or the tabs will not appear

##### Tab 2
Same rule as above

mark-clements

  ##############################################################################
  # tidyr
  ##############################################################################
  library(tidyr)
  #####################################
  # gather columns into key-value pairs
  #####################################
  # get first observation for each Species in iris data -- base R
  mini_iris <- iris[c(1, 51, 101), ]
  

mark-clements

  ################################################################################
  # File includes / common script headers
  ################################################################################
  library(data.table)

  # setwd à l'emplacement du fichier
  root <- setwd(dirname(rstudioapi::getActiveDocumentContext()$path))
  ################################################################################
  # Vignette, excerpts
  ################################################################################