Inferring City-wide Travel Patterns Using Mobile Phone Data (Call Detail Records): Case Study of Tehran

In this study, we utilized mobile phone call detail records (CDR) to extract individuals’ travel patterns in Tehran, trying to demonstrate the potential and opportunities of employing CDR for studying mobility patterns compared to traditional household surveys. In this regard, the raw CDR is firstly refined and preprocessed, and the initial dataset (containing about 1.6 billion records of 12 million people) is reduced to a sample of about 400,000 users with 500 million records. Then, using a series of heuristic and learning-based clustering algorithms, we identified each individual’s activity locations and inferred their trip purposes. By identifying the origin and destination of trips for individuals, the O-D matrices of the CDR sample are extracted. The sample O-D matrices are then multiplied by expansion factors to expand the sample to the population. We observed that these matrices are satisfactorily consistent with the travel patterns of people in Tehran extracted from Tehran’s Transportation Masterplan studies. Validating the results against the previous surveys shows that the proposed methodology is applicable in transportation and urban planning.