Aggregate Data in ThingSpeak Channel

This example shows how to aggregate data to a lower time resolution in a ThingSpeak™ channel to remove irregularity. Irregularity in a data can be caused due to several factors such as event driven sensing, malfunctioning of sensors, or network latencies.

Read Data

ThingSpeak channel 22641 contains tide and weather data measured once a minute at Ockway Bay, Cape Cod. Field 2 of the channel contains air temperature data. Read the air temperature data for the past 3 hours from channel 22641 using the thingSpeakRead function.

datetimeStop = dateshift(datetime('now'),'start','hour');
datetimeStart = dateshift(datetime('now'),'start','hour') - hours(3);

data = thingSpeakRead(22641,'DateRange',[datetimeStart,datetimeStop],...
    'Fields',2,'outputFormat','timetable');

Aggregate the Data

Data is measured once every minute. However, due to network latency associated with the measurement system, the actual timestamps can be greater than or less than a minute apart. Further, for the application of interest, the frequency of data measured every minute is high. Data at an hourly time resolution is sufficient. You can use the retime function to aggregate the data for each hour to a single value. You can use the maximum value for each hour to aggregate the data. Preview the first four values of the data with head.

dataHourly = retime(data,'hourly','max');
head(dataHourly,4)

ans =

  3×1 timetable

         Timestamps         AirTemperatureC
    ____________________    _______________

    03-Jan-2019 14:00:00          7.5      
    03-Jan-2019 15:00:00          6.9      
    03-Jan-2019 16:00:00          6.4      

Send Data to ThingSpeak

Change the channelID and the writeAPIKey to send data to your channel

channelID=17504;
writeAPIKey='23ZLGOBBU9TWHG2H';
thingSpeakWrite(channelID,data,'writeKey',writeAPIKey);