按鸟类个体及匹配时间分组计算不同时段平均体温差值

Got it, let's work through how to calculate those hourly average temperature differences for each bird across the different periods (PreI, DayI, PostI). Here's a step-by-step approach using pandas, which is perfect for this kind of tabular data manipulation:

First, we need to get the average Tb for each combination of Bird_ID, Hour, and Tb_Period. This groups the data so we're working with the hourly mean for each bird in each period—exactly what we need for the time-point comparisons.

import pandas as pd

# Calculate hourly mean Tb for each Bird_ID + Hour + Tb_Period combo
hourly_mean_tb = df.groupby(['Bird_ID', 'Hour', 'Tb_Period'])['Tb'].mean().reset_index()

Next, we'll pivot the data so each Tb_Period becomes a column. This puts all the hourly averages for a single bird and hour on one row, making it trivial to compute differences between periods.

# Pivot to wide format: Bird_ID + Hour as index, Tb_Period as columns
wide_tb = hourly_mean_tb.pivot(index=['Bird_ID', 'Hour'], columns='Tb_Period', values='Tb').reset_index()

Your output will look something like this (example for Bird 3282 at hour 11):

Now we can calculate the specific comparisons you need—like 09:00 DayI - 09:00 PreI or 10:00 PostI - 10:00 PreI. We'll add these as new columns to our wide dataframe.

# Calculate your desired period differences (customize these as needed)
wide_tb['DayI_minus_PreI'] = wide_tb['DayI'] - wide_tb['PreI']
wide_tb['PostI_minus_PreI'] = wide_tb['PostI'] - wide_tb['PreI']
# Add more if needed, e.g.:
# wide_tb['PostI_minus_DayI'] = wide_tb['PostI'] - wide_tb['DayI']

If you only care about certain hours (like 09:00 and 10:00), you can filter the dataframe to focus exclusively on those time points:

# Keep only rows for hours 9 and 10
target_hours = wide_tb[wide_tb['Hour'].isin([9, 10])]

• Double-check that your Hour column is formatted as an integer (e.g., 9 for 09:00, 10 for 10:00). If it's a string, convert it first with df['Hour'] = df['Hour'].astype(int).
• If some birds are missing data for a specific period/hour, you'll get NaN values in the difference columns. Handle these with fillna() (to replace with a default value) or dropna() (to remove incomplete rows) as needed.
• This setup is flexible—you can easily add more period comparisons later by extending the difference calculations.

内容的提问来源于stack exchange，提问作者 S.Bird