Resampling time series data in Pandas - Step-by-Step Execution

Concept Flow - Resampling time series data

Start with time series data

↓

Choose resampling frequency

↓

Apply resample() method

↓

Aggregate data (mean, sum, etc.)

↓

Get new resampled time series

↓

End

We start with time series data, pick a new time interval, resample using pandas, aggregate values, and get a new summarized time series.

Execution Sample

Pandas

import pandas as pd

# Create sample data
idx = pd.date_range('2024-01-01', periods=6, freq='H')
data = pd.Series([10, 20, 15, 30, 25, 40], index=idx)

# Resample hourly data to 3-hour intervals
resampled = data.resample('3H').mean()
print(resampled)

This code creates hourly data and resamples it to 3-hour intervals by averaging values.

Execution Table

Step	Action	Input Data	Resample Frequency	Aggregation	Output Data
1	Create hourly time series	[10,20,15,30,25,40]	Hourly (H)	None	2024-01-01 00:00:00 -> 10 2024-01-01 01:00:00 -> 20 2024-01-01 02:00:00 -> 15 2024-01-01 03:00:00 -> 30 2024-01-01 04:00:00 -> 25 2024-01-01 05:00:00 -> 40
2	Choose resample frequency	Hourly data	3 Hours (3H)	None	Preparing to group data into 3-hour bins
3	Group data into 3-hour bins	Hourly data	3H	None	Bin 1: 00:00-02:59 -> [10,20,15] Bin 2: 03:00-05:59 -> [30,25,40]
4	Aggregate each bin by mean	Bins	3H	Mean	Bin 1 mean: (10+20+15)/3 = 15 Bin 2 mean: (30+25+40)/3 = 31.67
5	Create new resampled series	Aggregated means	3H	Mean	2024-01-01 00:00:00 -> 15 2024-01-01 03:00:00 -> 31.67
6	Print resampled data	Resampled series	3H	Mean	Output: 2024-01-01 00:00:00 15.00 2024-01-01 03:00:00 31.67 Freq: 3H, dtype: float64

💡 All original data grouped and aggregated into 3-hour intervals, resampling complete.

Variable Tracker

Variable	Start	After Step 1	After Step 3	After Step 4	Final
data	None	[10,20,15,30,25,40] hourly indexed	[[10,20,15],[30,25,40]] grouped by 3H	[15, 31.67] means of groups	[15, 31.67] resampled series
resampled	None	None	None	None	[15.0, 31.67] with 3H freq

Key Moments - 3 Insights

Why does the resampled series have fewer rows than the original?

Because resampling groups multiple original time points into larger intervals (3 hours here), so fewer aggregated points appear as shown in execution_table rows 3 and 5.

What happens if we use sum instead of mean for aggregation?

The grouped values would be added instead of averaged, changing the output values in step 4 and 5 accordingly.

Why is the new index at 00:00 and 03:00 after resampling?

Because pandas labels each resampled bin by the start time of the interval, as shown in execution_table row 5.

Visual Quiz - 3 Questions

Look at the execution_table at step 4, what is the mean of the first 3-hour bin?

A 31.67

B 15

C 20

D 10

💡 Hint

Check the 'Output Data' column at step 4 for the first bin's mean.

At which step does the data get grouped into 3-hour bins?

A Step 4

B Step 2

C Step 3

D Step 5

💡 Hint

Look for the step describing grouping in the 'Action' column.

If we changed the resample frequency to '2H', how would the number of output rows change?

A More rows than with '3H'

B Fewer rows than with '3H'

C Same number of rows

D No rows

💡 Hint

Refer to variable_tracker 'data' row showing grouping by frequency.

Concept Snapshot

Resampling time series data with pandas:
- Use .resample('freq') on a time-indexed series or DataFrame
- 'freq' is new time interval (e.g., '3H' for 3 hours)
- Aggregate grouped data with mean(), sum(), etc.
- Result is a new time series with fewer or more points
- Index labels are interval start times

Full Transcript

This visual execution shows how pandas resamples time series data. We start with hourly data points. We pick a new frequency, here 3 hours. The data is grouped into 3-hour bins. Each bin's values are aggregated by mean. The output is a new series with fewer points, each representing the average over 3 hours. The index labels are the start times of each 3-hour interval. This process helps summarize or change the time scale of data easily.