7 Pandas Methods everybody should know

Lo and Behold, here i am back with the 7 most important methods in pandas that every data scientist should be aware of.

Note: I’ll be skipping over the basic methods like head() and info(), coz they are the best.

1) describe() - This method can be applied on both Data Frames and Series; its a quick way to breeze through the descriptive statistics of the columns involved in the dataset. - The 5 point summary of data can be a good starting point for getting an idea about the data distribution. The Five Numbers are Minimum, Q1(1st Quartile), Median, Q3(3rd Quartile) and Maximum - Along with that, the method also provides the frequency(count),mean and standard deviation to better understand the spread. - For object(string) columns; the above method provides us with the count, number of unique objects, the most frequent element and also its frequency.

```python
import pandas as pd 
df=pd.DataFrame({'Age':[14,35,25,36,19],'Height':[144,157,178,181,171]})
df.describe()
#output
#>>> df.describe()
#         Age      Height
#count   5.000000    5.000000
#mean   25.800000  166.200000
#std     9.679876   15.482248
#min    14.000000  144.000000
#25%    19.000000  157.000000
#50%    25.000000  171.000000
#75%    35.000000  178.000000
#max    36.000000  181.000000 
``` 2) `loc` & `iloc`
Both these methods are an absolute must-knows. They are used to retrieve subsets of the data. 
loc :- label based indexing
iloc:- integer based indexing

 An example would clarify the difference between them well

	df=pd.DataFrame({'Age':[14,35,25,36,19],'Height':[144,157,178,181,171]})
	
	# Let me write the code to extract the first column using both the methods 
	df.iloc[:,0]  
	df.loc[:,'Age'] 

3) groupby Its the go-to method for any data aggregation task, if you want to understand the inter-relationship between features. ```python #Lets up the ante by making the dataframe abit more diverse df=pd.DataFrame({‘Age’:[14,35,25,36,19],’Height’:[144,157,178,181,171],’Gender’:[‘Male’,’Female’,’Female’,’Male’,’Female’]})

# Suppose we are interested in the average height and age for both males and female, fret not my friend, Here comes groupby to the rescue. 
df.groupby(by='Gender')['Height'].mean() # average height of both the classes
# similarly 
df.groupby(by='Gender')['Age'].mean() # average age of both the classes ```

4) value_counts If you are working with a categorical feature and would want to find the number of occurrences of each class, then value_counts is the method you should be using. Let me run you through a quick example, ```python df=pd.DataFrame({‘Age’:[14,35,25,36,19],’Height’:[144,157,178,181,171],’Gender’:[‘Male’,’Female’,’Female’,’Male’,’Female’]})

#to find the number males and females in the data, we could use 
df['Gender'].value_counts() ```

5) pivot_table - It’s the same pivot table which used to give us nightmares in Excel. But Pandas makes it a tad bit easy is what i believe. - With pivot table, you can build super complex aggregations, sometimes the code just bends my mind. - Here is the syntax for the curious few of you: pandas.pivot_table(_data_, _values=None_, _index=None_, _columns=None_, _aggfunc='mean'_, _fill_value=None_, _margins=False_, _dropna=True_, _margins_name='All'_, _observed=<no_default>_, _sort=True_)

```python
df=pd.DataFrame({'Age':[14,35,25,36,19],'Height':[144,157,178,181,171],'Gender':['Male','Female','Female','Male','Female']})

pd.pivot_table(df,values=['Age','Height'],index='Gender',aggfunc='mean')
#Output
#Gender
#Female  26.333333  168.666667
#Male    25.000000  162.500000
```  6) `apply`
-  If you wanna apply a function to all the entries of a column, then apply will make your life simple. 
- Just define the operation as a regular function and then just *apply* it.
- If you want to go more pythonic, then try lambda functions(they just make the code look more sleek)
```python
def lower(s: str)-> str:
	return s.lower()

df=pd.DataFrame({'Age':[14,35,25,36,19],'Height':[144,157,178,181,171],'Gender':['Male','Female','Female','Male','Female']})

df['Gender']=df['Gender'].apply(lower)
``` 7) `replace`
- The name of the method simply explains what it does, replaces the values you want to be replaced with the given replacement. 
```python
df=pd.DataFrame({'Age':[24,35,25,36,29],'Height':[144,157,178,181,171],'Gender':['Male','Female','Female','Male','Female'],'Marital Status':['Not Interested','Alone','Married','Depressed','Single']})

df.replace(to_replace=['Not Interested','Alone','Depressed'],value='Single')
```

Pandas is a great library and has plenty more mind boggling methods which can make our lives simpler.

Head over to the official Pandasdocumentation to read more about them.

Happy Reading!

Enjoy Reading This Article?