The data
What does the data collected by the wifi estimator look like?
At the core, we track presences. They look like this:
2697866 2022-11-04 00:16:52.000 -0400 2022-11-04 02:02:51.000 -0400 KY0069-002
2697865 2022-11-03 20:48:53.000 -0400 2022-11-04 02:58:51.000 -0400 KY0069-002
2697864 2022-11-04 01:00:51.000 -0400 2022-11-04 02:52:51.000 -0400 KY0069-002
2697863 2022-11-04 02:48:51.000 -0400 2022-11-04 02:54:51.000 -0400 KY0069-002
2697862 2022-11-03 20:48:53.000 -0400 2022-11-04 02:58:51.000 -0400 KY0069-002
2697861 2022-11-03 21:10:52.000 -0400 2022-11-03 21:10:52.000 -0400 KY0069-002
The first number is an index; it has no meaning, other than to be a counter for the data row.
The second number is a start time. So, the first row is 12:16AM on the 4th of November.
The third number is an end time. This is 2:02AM on the 4th of November.
The last field is the unique id of the library; it is entered as part of the sensor setup process. This data is not actually from Berea, Kentucky (KY0069-002 is the Berea branch of the Madison County Public Library system). We're just fans of Berea, KY, and used it as a test value.
We can put that data into a simple histogram, where we look at how many devices were seen throughout a given day.
We might also want to see a whole week's worth of data.
In this example, you might wonder what the zeros are. We some of the devices in our initial pilot were overheating, and shutting down for periods of time; we've worked to correct that in our next phase.
Yes! The data is in the public domain.
We won't go into that on this page, but it is worth knowing that we provide both a browser for the data (which the images above were pulled from) as well as a way for you to access the data with code (an API, as the computer people like to say).
Examples for how to access and manipulate that data are provided later in this documentation.