Work With Mobile Phone Data

- - posted in study

MOBILE PHONE DATA allows researchers to study individuals’ movement for a long period. Therefore, in recent years, it becomes killer mace in travel behavior analysis and human mobility research. However, Indicating a dataset as mobile phone data is quite ambiguous. In general, when we mention mobile phone data, it could be two different things.

1. Data collected by mobile phones
This data is usually collected by LBS (location-based service) providers, like Google, Uber, Nike. Sometimes it is also collected for research programs, like program in Singapore and 20 DAY STRANGER. Data collected by mobile phones can be also divided by locating methods listed in the following:

Method Mechanism Positional Accuracy Applying Conditions Power Consumption
GPS locating Caculate the location based on the GPS signal 10m Outdoor, no big cover
Large, in general will deplete the battery in 3 hours
WIFI locating Query the location based on the wifi’s MAC address 30-200m (Depend on the service area of the wireless transmitter) Has wifi, can access the internet
Small
Base station locating Query the location based on the ID of base stations 300m-2000m (Depend on the service area of the base station) Has signal, can access the internet
Small


Data collected by GPS or WIFI has very high locating precision. The following video shows a phone-collected trajectory of a student who moves from one campus of Tongji University to another. One can easily tell how did he/she move since the data has remarkable precision.

However, in most cases, it is hard to obtain large-scale phone-collected data. For researchers, it is hard to initiate a survey to obtain people’s data in large scale. For big companies, they have difficulty in obtaining the whole picture of human’s travel: Uber’s data only records the usage of uber’s car, Nike+ only records cycling or running tracks, most map apps records data only when people open the map.

2. Data collected by base stations
This data is collected by base stations and it is collected for billing and operational purposes. It contains no locational information but can be used to estimate the user’s location based on the location of the base station.

Since its passive collecting nature, this data can serve a human mobility research in urban level or even larger. The above figure shows one-day temporal-spatial trajectories of all mobile users in Shanghai. One can find 1.Data collected by base stations can roughly reflect one’s travel trajectories; 2. These “colorful lines” show some certain patterns, like they converge in the daytime and diverse in the night.

To conclude, mobile phone data is amazing and when seeing a “mobile phone data”, we need further information to decide what it really is.