I have data sets that represent the temperature of an area over the course of the day. I have data for several months, but some of these data are missing for hardware reasons. Data may be missing for several hours, or even several hours at a time.
I therefore had to add NaN values to my dataset in order to better represent them. I've filled in the missing data for the hours in a day, but not for several days of empty data.
I'd like to display temperature values as a function of time and for an entire month of data, using a curve.
I'd like to display only non-empty data, without matplotlib linking unrelated data.
The problem is that in this example, some data are linked and others are not, even though there are several days of missing data. We can see data linked around 2023/10/09, while between 2023/10/21 and 2023/10/29 the data is not linked despite the fairly substantial missing data.
To solve this problem, I've already tried displaying the data day by day to avoid matplotlib linking the data together, but unfortunately with this solution I end up with as many legends as days of data. here's the code that gives me the above result:
"""
Create and display the temperature distribution plot (curves) for a given month.
Parameters:
month_year_data (pandas.DataFrame): DataFrame containing hourly temperature data for the month.
month_name (str): Name of the month for labeling the plot.
year (int): Year for labeling the plot.
Returns:
None
"""
# Create a single figure
fig, ax = plt.subplots(figsize=(20, 10))
# Plot temperature distribution for each zone
zones = ['Far range', 'Mid range', 'Near range']
colors = ['blue', 'green', 'red', 'black'] # Define colors for each zone
for i, zone in enumerate(zones):
ax.plot(month_year_data.index, month_year_data[zone], label=zone, color=colors[i], linestyle='-')
# Set labels for x and y axes
ax.set_xlabel('Time')
ax.set_ylabel('Temperature (°C)')
# Set title for the plot
ax.set_title(f'Temperature as a function of time by Hour - {month_name}-{year}')
# Rotate x-axis labels for better readability
plt.xticks(rotation=45)
# Add legend
ax.legend()
I've filled in the missing data for the hours in a day, but not for several days of empty data.
Matplotlib will link consecutive points if there is no NaN in between, irrespective of the actual time difference that the points correspond to. I suggest you divide the whole interval into hours which are equally spaced, and fill the array with the data you have, and the missing data with NaN. You might need to add rows to the dataframe, if there is no entry for a particular date, for example. You might want to look at this answer and this one.
I see your data is already indexed by date/time. Try:
rng = pd.date_range(month_year_data.index.min(), month_year_data.index.max(), freq='H')
month_year_data_filled = month_year_data.reindex(rng)
Missing values should be automatically filled NaN, as it is the default.