This data is useless and no conclusions can be drawn from it
A) no controls to ensure it's a statistically valid sample. Users who volunteer to be measured, and in this case actually take the trouble to download and activate a metering app, are almost certainly not your average user. Also the sample size is way too small.
B) no report of the skews. By all reports, cell data usage is a highly skewed curve. Average (mean) usage is not very meaningful. If you want a single number, median is more meaningful. Most useful would be a set of percentage breakouts. E.g., 88% of users use less than 1.2 GB/mo.
But, that would not feed the "caps are bad and users are getting overcharged" narrative. So, the data gets reported in a misleading way to support the desired conclusion.