Self notes

• The analysis itself is secondary. The base data column, the key facts you collect are much more important.
• EG: While predicting the sales of a list of stores in a city– Count of items, in store, sqft area of store, outlet type, city type in store

Preparations

• Deal with null values (Substitute them with mean)
• Convert catagorial columns into 1-hot
  • from sklearn.preprocessing import LabelEncoder

Definitions: Target variable

• The varibale you want to successfully predict after your analysis ID columns
• Remove ‘ID’ type columns from your DF as they do not add any value to the predictive model

Process

• PreProcess
  • Bring all columns to same same value count
• Explore
  • Displot Each column(see value distribution)

Explore

Counts

In this phase, we want to make sure there is uniform distribution of values Displot

• Here we want to see how the values in the column are distributed
• It’s skew (Center, Left, Right)
• If Left / Right skew then apply log transformation
  • np.log(1+df['sales']) Countplot
• sns.countplot()

Correlation

• sns.heatmap(df.corr())