Create Timetables

This example shows how to create a timetable, combine timetables, and adjust the data from multiple timetables to a common time vector. The common time vector can contain the times from either or both timetables, or it can be an entirely new time vector that you specify. The example shows how to compute and display a daily mean for weather measurements contained in different timetables.

A timetable is a type of table that associates a time with each row. A timetable can store column-oriented data variables that have different data types and sizes, so long as each variable has the same number of rows. In addition, timetables provide time-specific functions to combine, subscript into, and adjust their data.

Import Timetables from Files

Load air quality data and weather measurements into two different timetables. The dates of the measurements range from November 15, 2015, to November 19, 2015. The air quality data come from a sensor inside a building, while the weather measurements come from sensors outside.

Read the air quality data from a table with the readtimetable function. The output is a timetable.

indoors = readtimetable('indoors.csv');

You also can create a timetable from an M-by-N array with the array2timetable function, or from workspace variables with the timetable function.

Display the first five rows of indoors. Each row of the timetable has a time that labels that row of data.

indoors(1:5,:)

ans=5×2 timetable
           Time            Humidity    AirQuality
    ___________________    ________    __________

    2015-11-15 00:00:24       36           80    
    2015-11-15 01:13:35       36           80    
    2015-11-15 02:26:47       37           79    
    2015-11-15 03:39:59       37           82    
    2015-11-15 04:53:11       36           80    

Load the timetable with weather measurements. Display the first five rows of outdoors.

load outdoors
outdoors(1:5,:)

ans=5×3 timetable
           Time            Humidity    TemperatureF    PressureHg
    ___________________    ________    ____________    __________

    2015-11-15 00:00:24        49          51.3          29.61   
    2015-11-15 01:30:24      48.9          51.5          29.61   
    2015-11-15 03:00:24      48.9          51.5          29.61   
    2015-11-15 04:30:24      48.8          51.5          29.61   
    2015-11-15 06:00:24      48.7          51.5           29.6   

Synchronize Timetables

The timetables, indoors and outdoors, contain different measurements taken inside and outside a building at different times. Combine all the data into one timetable with the synchronize function.

tt = synchronize(indoors,outdoors);
tt(1:5,:)

ans=5×5 timetable
           Time            Humidity_indoors    AirQuality    Humidity_outdoors    TemperatureF    PressureHg
    ___________________    ________________    __________    _________________    ____________    __________

    2015-11-15 00:00:24           36               80                49               51.3          29.61   
    2015-11-15 01:13:35           36               80               NaN                NaN            NaN   
    2015-11-15 01:30:24          NaN              NaN              48.9               51.5          29.61   
    2015-11-15 02:26:47           37               79               NaN                NaN            NaN   
    2015-11-15 03:00:24          NaN              NaN              48.9               51.5          29.61   

The output timetable, tt contains all the times from both timetables. synchronize puts a missing data indicator where there are no data values to place in tt. When both input timetables have a variable with the same name, such as Humidity, synchronize renames both variables and adds both to the output timetable.

Synchronize the timetables again, and this time fill in missing data values with linear interpolation.

ttLinear = synchronize(indoors,outdoors,'union','linear');
ttLinear(1:5,:)

ans=5×5 timetable
           Time            Humidity_indoors    AirQuality    Humidity_outdoors    TemperatureF    PressureHg
    ___________________    ________________    __________    _________________    ____________    __________

    2015-11-15 00:00:24            36                80               49               51.3         29.61   
    2015-11-15 01:13:35            36                80           48.919             51.463         29.61   
    2015-11-15 01:30:24         36.23             79.77             48.9               51.5         29.61   
    2015-11-15 02:26:47            37                79             48.9               51.5         29.61   
    2015-11-15 03:00:24            37            80.378             48.9               51.5         29.61   

Adjust Data in One Timetable

You also can adjust the data in a single timetable to a new time vector. Calculate the means of the variables in ttLinear over six-hour intervals with the retime function. If any rows have NaN values after you adjust the data, remove them with the rmmissing function.

tv = [datetime(2015,11,15):hours(6):datetime(2015,11,18)];
ttHourly = retime(ttLinear,tv,'mean');
ttHourly = rmmissing(ttHourly);

Plot Timetable Data

Normalize the data in ttHourly to the mean for each variable in the timetable. Plot the mean daily values of these measurements. You can use the Variables property of a timetable to access the variables. ttHourly.Variables returns the same variables as ttHourly{:,:}.

ttMeanVars = ttHourly.Variables./mean(ttHourly.Variables);
plot(ttHourly.Time,ttMeanVars);
legend(ttHourly.Properties.VariableNames,'Interpreter','none');
xlabel('Time');
ylabel('Normalized Weather Measurements');
title('Mean Daily Weather Trends');