It's easy to record something cheaply, but as other commenters already wrote, it's hard to recognize what causes the "average" counts in any time period. The author of the article measured the variations dependent of the time of day 17 years ago and it seems he wasn't able to explain them since. At the time he used a 486 machine and a Geiger counter connected to the serial port, today something like Raspberry Pi would be more convenient for the computer side and less power hungry.
He published what he recorded, so before doing anything, check his data to see an example of the results that you'd have in such a setup.
http://www.fourmilab.ch/documents/cosmic/
It's easy to record something cheaply, but as other commenters already wrote, it's hard to recognize what causes the "average" counts in any time period. The author of the article measured the variations dependent of the time of day 17 years ago and it seems he wasn't able to explain them since. At the time he used a 486 machine and a Geiger counter connected to the serial port, today something like Raspberry Pi would be more convenient for the computer side and less power hungry.
He published what he recorded, so before doing anything, check his data to see an example of the results that you'd have in such a setup.