Methods to calculate data flow To & From Imp over a fixed period

methods to calculate data flow To & From Imp over a fixed period. i am thinking of placing an imp in a remote position. Power is available, but no wifi, so i thought about a dedicated powered/battery prepaid module. i can get a unit for about $15 per month, with 1 gigabyte included data. I know the imp would not use that amount, but as an exercise, i would like to try and calculate how much data would be consumed. I have an imp 01 , with a ds18b20 on uart12, and 1 ds18b20 on uart57, sending the temperatures to thingspeak, at the rate of 1 reading per 3 minutes for each sensor, so ( 20/ hour, 480/day, ) X 2, = 960 readings total per day. The expected temperature range is about -5C to about 35C . Thingspeak logs the 2 temperatures as seperate graphs, and also combines them into 1 nice graph, indoor/outdoor temperatures What are the implications of just running the unit with code, (send data, wakeup, send data wakeup etc). I am not logging data to the imp server, as i have a 16/2 line LCD connected, so can see the readings on the LCD display. Any good tools/ software available to see the data flow up/down for a PC. I have no idea how much data such a setup would consume, in a day or month, any input from forum members welcome. TOM.
I had the device logging last night, but it is tethtered to my mobile for wifi, so when i get a more permanent setup, i could post the imp code and the thingspeak link if anyone is interested. 0.5 C outside last night at 3 am. Sydney Australia, (middle of winter).

Have you considered a solar panel?

A solar setup, as described in the link is a great idea for portable/remote sensors, providing you have done all your calculations. In the situation i was considering, the primary concern i had was about no wifi being within range, therefore how to get the data out. That was where my thoughts turned to “portable” wifi access points, and what is available. Thinking more about it, a budget mobile phone could function purely as an access point, as long as it was kept charged, but it would need to be on all the time with the wifi enabled, and so would draw a fair amount off current, even if you are only acessing the wifi for a few seconds to send the data, every 5 or 6 minutes. Maybe a “cellular imp006” might be a viable alternative if there was ever going to be such a thing, that would certainly open up a whole new range of options for passing data. I can see that the benefits of what IOT devices purport to do in enabling lots of devices connected to the cloud are relying on there always being a method to connect being available, but there must be many situation where there is no wifi available. Relatively recently there appears to be a lot of focus and attention on LORA and SIGFOX, to fill the gap where low data rates at low frequency send intervals are being employed in remote situations, for sensor data, ( temperature, humidity, environmental data, etc). Who knows how all these connected devices will be employing, not only in the short term, but say in 10 years from now.