I need a dataset for classifying users based on distance and time taken ( with delay spread). So I'll be happy if I get some suggestions or existing datasets.
The carriers keep data that shows the estimated distance a user is located from a cell tower at regular intervals, but I have evaluated this data and it is not reliable. So be careful if you are offered RTT (Round Trip Time) or NELOS data.
The phone companies have been under pressure for years to improve the 911 (Emergency calls) system. Today 911 uses GPS and the standard is that the phone companies must be able to locate a caller within 50 meters 65% of the time. It works fairly well, but is not useful to determine if a caller is on the 4th floor of a building, nor is it generally accurate inside buildings. So, efforts to use RTT to calculate the distance from a tower have not moved forward. The data is likely useful for measuring the estimated power levels needed to complete phone calls based on the perceived distance.
I regularly receive this RTT data and when I do analysis of cell phone records. I have evaluated these records and usually find that distance from tower measurements that are seconds or minutes apart, so resulting distances that are not credible. (For example, two measurements over 20 seconds show that the device has moved 6 Km. In other cases, a device is known to be stationary for several hours but the measurements shows it's location to move about.)
Some of the carriers add a "confidence" parameter to each reading to let you know how accurate the location measurement may be (based on a number of factors including the number of satellites used for the measurement.)
Since the distance to the tower estimates are based on the speed of the radio signal, if you are off by 1 millionth of second the error is .18 of a mile or so.
In the end, GPS will not be the technology of choice for determining the location of a caller in the future. By 2025 we will see other technologies used for location, especially indoors.
To finally answer your question: You will find this data but it may not be helpful. If you find something more reliable, please let us know.