Python Moving Average
Creating a moving average is a fundamental part of data analysis. You can easily create moving averages with Python data manipulation package. Pandas has a great function that will allow you to quickly produce a moving average based on the window you define. This window can be defined by the periods or the rows of data.
Pandas ROLLING() function:
The rolling function allows you aggregate over a defined number of rows. If your rows are based on a day level of granularity, you will be aggregating over the day levels.
There is a window parameter to allow you to specify the span of the your aggregation. Here is an example of rolling average for a 3 day window average if your data is at the day granularity.
ROLLING(WINDOW=3).MEAN()
Applying this function to a column would look similar to the below:
df['Column'].rolling(window=3).mean()
Let's look at an example of this using our avocado dataset.
Requirements
- datetime index
These function works off datetime index. So, you will need to setup the data below:
#load to the depencies
import pandas as pd
df = pd.read_csv('avocado.csv',index_col='Date',parse_dates=True)
%matplotlib inline
#Limit the datas set and aggregate the to monthly. This is not necessary but for this data set it easier to visualize.
df.head()
df =df.loc['2015-12-12':]
df = df.resample('M').sum()
#create and visualize the rolling average
df['Total Bags'].plot()
df['Total Bags'].rolling(window=3).mean().plot()
