Exploring data cards

Name 4 ways to do EDA (Exploratory Data Analysis)

• compare columns against each other (feature variables)

• compare variables against your target (what we want to predict)

• Understand the data dictionary

• talk with SME's so that, you become an SME on this data

EDA - Exploratory data analysis: 4 thing s u need to know about the data

• what kind of data is this (catagorical, numeric etc.)? What tool and which models make the most sense to use?

• Are their outliers?

• How will missing data be dealt with?

• Which features (variables) should be used (kept or discarded)?

How to use code to EDA the target variable

dataframe['target'].value_counts()

# of the 303 patients (each is a row of data)

# 165 have heart disease vs. 138 do not have heart disease

# Roughly 50/50 so our data has an equal split roughly 150 in each target group - balance data

Use matplotlib.pyplot as plt to visualize the data given the previous variables. dataframe = pd.read_csv("data/heart-disease.csv") and %matplotlib inline

dataframe['target'].value_counts()

# of the 303 patients (each is a row of data)

# 165 have heart disease vs. 138 do not have heart disease

# Roughly 50/50 so our data has an equal split roughly 150 in each target group - balance data