Revision history [back]
 | 1 | initial version |
One way to reduce the size of a DataFrame through downsampling is by randomly selecting a subset of rows from the original DataFrame. This can be done using the sample() method in pandas.
For example, if we want to downsample our DataFrame df to 50% of its original size, we can use the following code:
df_downsampled = df.sample(frac=0.5)
This will randomly select 50% of the rows from df and return them in a new DataFrame called df_downsampled.
Another way to downsample a DataFrame is by aggregating, i.e., grouping rows by a certain column or set of columns and reducing the number of rows by performing some calculation on the groups, such as taking the mean or sum of the values in each group. This approach is useful if we want to summarize the data in some way, but it may not be appropriate if we want to preserve the individual observations in the DataFrame.