A-TPN: User trajectory prediction method based on mobile communication CDR data

Examining potentials and practical constraints of mobile phone data for improving transport planning in developing countries

ABSTRACTIn recent years, call detail records (CDRs) have been widely used in human mobility research. Although CDRs are originally collected for billing purposes, the vast amount of digital footprints generated by calling and texting activities provide useful insights into population movement. However, can we fully trust CDRs given the uneven distribution of people’s phone communication activities in space and time? In this article, we investigate this issue using a mobile phone location dataset collected from over one million subscribers in Shanghai, China. It includes CDRs (~27%) plus other cellphone-related logs (e.g., tower pings, cellular handovers) generated in a workday. We extract all CDRs into a separate dataset in order to compare human mobility patterns derived from CDRs vs. from the complete dataset. From an individual perspective, the effectiveness of CDRs in estimating three frequently used mobility indicators is evaluated. We find that CDRs tend to underestimate the total travel distance and the movement entropy, while they can provide a good estimate to the radius of gyration. In addition, we observe that the level of deviation is related to the ratio of CDRs in an individual’s trajectory. From a collective perspective, we compare the outcomes of these two datasets in terms of the distance decay effect and urban community detection. The major differences are closely related to the habit of mobile phone usage in space and time. We believe that the event-triggered nature of CDRs does introduce a certain degree of bias in human mobility research and we suggest that researchers use caution to interpret results derived from CDR data.

Machine learning techniques have been used to predict customer behavior in telecommunications industry. Typically, several data sources, including historical usage, billing, payment, network, customer service, and demographic data, can be used in a predictive model. However, in some cases, e.g., in the prepaid customer segment, there is often little data available except for the call detail record (CDR) data. In this paper, we tackle this challenging problem, using significantly delayed CDR data as the primary data source to predict customer behavior. We extract calling links, i.e., who called whom, from the CDR data, and propose several distance measures based on calling links. We demonstrate that, by using information derived from the calling links alone as inputs to a neural network model, an acceptable accuracy for predicting churn (customer switching from one service provider to another) can be achieved. Calling links can also be used to identify calling communities, which may be used for targeted marketing campaigns and help predict acceptance of marketing offers.

Mobile positioning is a key element in many geolocation applications and research fields about human mobility patterns, location-based services, targeted marketing, urban mobility, public health, and transport planning. The commonly used data for understanding the large scale mobility patterns are mobile network data or Call Detail Records (CDRs). However, CDR data has two major drawbacks: temporal and spatial uncertainties and sparseness. Although the first problem is widely covered by trajectory reconstruction techniques, the second problem still remains challenging. Hence, in this paper, we propose the adaptation of a new method based on a particle filter algorithm for mobile positioning and trajectory reconstruction. Our goal is to evaluate if this nonlinear method can out-perform the existent linear methods like Switching Kalman Filter. Therefore, the model performance and the effects of the parameters on accuracy were evaluated in controlled experimental settings. Additionally, the experiments were performed on a real dataset and compared with the results achieved by a linear approach.

The interior of urban rail transit system has always been the blind area of traditional passenger flow monitoring methods. The call detail record (CDR) data is a kind of data containing passenger travel information with the advantages of wide coverage and no dead angle. It can well complement the shortcoming that traditional passenger flow monitoring method has blind areas. This paper proposes an approach of urban rail passenger flow monitoring based on CDR data which can directly monitor passenger flow in metro station and passenger flow on different routes. It makes up the defect that previous research can only estimate passenger flow by models. Experiment with actual data, the one-week passenger flow of Beijing Metro Line 2 is monitored. It shows that the characteristics of passenger flow are closely related to the nature of land use and passengers are more likely to be affected by transfer times when transferring. The experimental results match reality and verify the feasibility of the algorithm.

With the increasing use of mobile devices, now it is possible to collect different data about the day-to-day activities of personal life of the user. Call Detail Record (CDR) is the available dataset at large-scale, as they are already constantly collected by the mobile operator mostly for billing purpose. By examining this data it is possible to analyze the activities of the people in urban areas and discover the human behavioral patterns of their daily life. These datasets can be used for many applications that vary from urban and transportation planning to predictive analytics of human behavior. In our research work, we have proposed a hierarchical analytical model where this CDR Dataset is used to find facts on the daily life activities of urban users in multiple layers. In our model, only the raw CDR data are used as the input in the initial layer and the outputs from each consecutive layer is used as new input combined with the original CDR data in the next layers to find more detailed facts, e.g., traffic density in different areas in working days and holidays. So, the output in each layer is dependent on the results of the previous layers. This model utilized the CDR Dataset of one month collected from the Dhaka city, which is one of the most densely populated cities of the world. So, our main focus of this research work is to explore the usability of these types of dataset for innovative applications, such as urban planning, traffic monitoring and prediction, in a fashion more appropriate for densely populated areas of developing countries.

Research on the detection of land use has been a challenge for urban sensing and is of great significance to smart city construction. With the popularity of smart phones and the unprecedented growth of mobile phone data, there emerge numerous studies on land use detection with CDR (Call Detail Records), SMS (Short Messaging Service) and mobile traffic data. This is also a combination of urban sensing based on mobile big data and smart city construction. In this paper, we propose a methodology based on existing studies on land use detection with CDR, text data, and mobile traffic data. We propose a new traffic signature that combines the advantages of the above studies and apply it to a real-world dataset collected in northeast of China. We choose the POI (point of interest) data as ground truth data that the obtained land use types are compared with. We find that our method can accurately detect the POI collections of specific types, which proves the effectiveness of our method. After extracting base components of land use types, an NMF (non-negative matrix factorization) based method is applied to explore the mix of land use types in urban areas. Then, we provide reasonable explanations for different areas with mixed land use types. The result of proposed method is evaluated from both components perspective and urban area perspective.

Human behavior is often affected by sudden emergency directly or indirectly. In this paper, we focus on two groups of people under different types of medical emergency calls. Based on the quantitative analysis of calling volume and call interval distribution before and after the emergency call, we find that during most types of emergencies, calling volume will increase after the emergency call made and maintain a high level in the next several hours. Medical emergency makes influence on call volumes, but not affects power-law distribution in human communication behavior. From studying entropy of call detail records, we found that during the medical emergency, the richness of outgoing call numbers increases and falls after that. In this paper, we do quantitative analysis of temporal patterns of medical emergency call based on the combination of medical emergency call data and call detail record data, which not only study on the characteristics of human communication behavior in emergency situations, but also provide data basis for scientific rational allocation of aid resources and effective emergency measures.

Call Detail Records (CDRs) have been widely used in the last decades for studying different aspects of human mobility. The accuracy of CDRs strongly depends on the user-network interaction frequency: hence, the temporal and spatial sparsity that typically characterize CDR can introduce a bias in the mobility analysis. In this paper, we evaluate the bias induced by the use of CDRs for inferring important locations of mobile subscribers, as well as their complete trajectories. Besides, we propose a novel technique for estimating real human trajectories from sparse CDRs. Compared to previous solutions in the literature, our proposed technique reduces the error between real and estimated human trajectories and at the same time shortens the temporal period where users' locations remain undefined.

Digital technologies have recently become more advanced, allowing for the development of social networking sites and applications. Despite these advancements, phone calls and text messages still make up the largest proportion of mobile data usage. It is possible to study human communication behaviors and mobility patterns using the useful information that mobile phone data provide. Specifically, the digital traces left by the large number of mobile devices provide important information that facilitates a deeper understanding of human behavior and mobility configurations for researchers in various fields, such as criminology, urban sensing, transportation planning, and healthcare. Mobile phone data record significant spatiotemporal (i.e., geospatial and time-related data) and communication (i.e., call) information. These can be used to achieve different research objectives and form the basis of various practical applications, including human mobility models based on spatiotemporal interactions, real-time identification of criminal activities, inference of friendship interactions, and density distribution estimation. The present research primarily reviews studies that have employed mobile phone data to investigate, assess, and predict human communication and mobility patterns in the context of crime prevention. These investigations have sought, for example, to detect suspicious activities, identify criminal networks, and predict crime, as well as understand human communication and mobility patterns in urban sensing applications. To achieve this, a systematic literature review was conducted on crime research studies that were published between 2014 and 2022 and listed in eight electronic databases. In this review, we evaluated the most advanced methods and techniques used in recent criminology applications based on mobile phone data and the benefits of using this information to predict crime and detect suspected criminals. The results of this literature review contribute to improving the existing understanding of where and how populations live and socialize and how to classify individuals based on their mobility patterns. The results show extraordinary growth in studies that utilized mobile phone data to study human mobility and movement patterns compared to studies that used the data to infer communication behaviors. This observation can be attributed to privacy concerns related to acquiring call detail records (CDRs). Additionally, most of the studies used census and survey data for data validation. The results show that social network analysis tools and techniques have been widely employed to detect criminal networks and urban communities. In addition, correlation analysis has been used to investigate spatial–temporal patterns of crime, and ambient population measures have a significant impact on crime rates.

This paper uses big data technologies to study base stations’ behaviors and activities and their predictability in mobile cellular networks. With new technologies quickly appearing, current cellular networks have become more larger, more heterogeneous, and more complex. This provides network managements and designs with larger challenges. How to use network big data to capture cellular network behavior and activity patterns and perform accurate predictions is recently one of main problems. To the end, first we exploit big data platform and technologies to analyze cellular network big data, i.e., Call Detail Records (CDRs). Our CDRs data set, which includes more than 1,000 cellular towers, more than million lines of CDRs, and several million users and sustains for more than 100 days, is collected from a national cellular network. Second, we propose our methodology to analyze these big data. The data pre-handling and cleaning approach is proposed to obtain the valuable big data sets for our further studies. The feature extraction and call predictability methods are presented to capture base stations’ behaviors and dissect their predictability. Third, based on our method, we perform the detailed activity pattern analysis, including call distributions, cross correlation features, call behavior patterns, and daily activities. The detailed analysis approaches are also proposed to dig out base stations’ activities. A series of findings are found and observed in the analysis process. Finally, a study case is proposed to validate the predictability of base stations’ behaviors and activities. Our studies demonstrates that big data technologies can indeed be utilized to effectively capture network behaviors and predict network activities so that they can help perform highly effective network managements.

The demand for mobile services is increasing exponentially and mobile telecommunication operators are facing new challenges such as bandwidth choking and resource allocation issues consequently. One of the most promising methods for overcoming these challenges is using Big Data Analytics on Call Detail Record (CDR). In this paper, a Long-Short Term Memory neural network is utilized for user activity prediction for the first time. The real CDR data is preprocessed and used for training the neural network. The accuracy of this prediction is evaluated in normal values and anomalies. The proposed system can predict user activity accurately using a limited amount of training data and its performance in anomalous behaviors is promising. The networks ability to be generalized is also evaluated with the cross-dataset test. The neural network is trained with one cells data and predict another cells activity properly. The proposed system is a step forward toward designing a practical and efficient CDR anomaly prediction system.

The ubiquitous massive mobile phone data generation presents new opportunities to determine the requirements of transportation, disaster management and public health care systems. Currently, data from mobile phone records can help in identifying the location of the users while they are making trips. Generally, this estimation is achieved using traditional data collection methods; however, these methods are difficult to apply in developing countries with rapidly growing cities owing to the high population and limitation in conducting a survey. Call detail records (CDRs) are used as base data because they are valuable data sources and can reduce the cost and time limitations. The aim of this study is to estimate origin-destination (OD) trips from each zone by using the CDRs. The OD trips are estimated by using the CDRs of one week taken from Myanmar Post and Telecommunication mobile operator for over 1.9 million users per day in Yangon, the economic center of Myanmar. The OD trips are estimated from CDRs based on the location of the base station in a limited time window and time frame. If the same mobile users is observed in two different the ones within the time limit, it is assumed that the mobile user is coming out from the first zone and the trips represents an originating trip. This trip would be the destination trip for zone where the mobile user enters. In this study, the originating (outgoing) and destination trips (incoming) from each township on a weekday and weekend are determined. These data are useful for infrastructure development and urban transportation planning.

Data-driven trajectory prediction with weather uncertainties: A Bayesian deep learning approach

Inferring dynamic origin-destination flows by transport mode using mobile phone data