It is easy to collect a great deal of data. However, understanding what the data means can be a challenge. For example, one of the many sensors in our house collects the amount of power that our HVAC system uses. Our air conditioner consumes the most power, by far, of any device in our house. The chart, above, shows the power our AC consumed recently over the course of several hours. Since it is the end of February and the weather is really nice, our AC does not run very often.
The chart at the top of this post is great as an overview of what is occurring. However, there is nothing here that can be used to, for example, identify trends and/or compel us to take action.
To take this data and actually use it for analysis and decision-making, we have a number of tools and techniques available. One of the very simplest is to create several Jupyter Notebooks. These will allow us to do some very elementary analysis of the data.
NuvIoT has a number of Python scripts that make this easy.
Authenticating with NuvIoT
You can create a very simple script to connect to our API services to get an authorization context to be used in additional calls.
Getting Your IDs
It is not very likely that you will remember all of the IDs for accessing data within your installation. We provide a script that will let you get lists of the different types of data that you may need as part of your data analysis.
Getting Data
Once you know the IDs of the devices you are looking for, getting the data for additional processing is very easy.
Elementary Time Series Analysis
The data presented in the above diagram might be interesting. But, just as with the chart we provided earlier in this post, it is difficult to use it to take action. There are many better approaches. We use simple range tests to demonstrate how we can start turning this raw data into meaningful events.
We used a decision tree that includes the known state information to determine how long the AC ran, how long the fan ran, and how long the AC was turned off. From this data, we can do some analysis. The length of time the AC ran gives us an idea of AC efficiency. The length of time the AC was off might help us understand how well-insulated the house is.
We can couple this data with other variables, such as the number of occupants, outside air temperature, sun cover, and attic temperature. At this point, we have enough information to make changes in our environment to reduce our electric bill or determine that our AC might not be operating as efficiently as it could have been.
To learn how Software Logistics and NuvIoT can help your organization better understand all the data produced by your sensors, please Contact Us.