“Our project started out of our belief in the power of data. We have already seen so many use cases involving commercial data. So, won’t it be better to bring experts from various fields together to solve social problems through data? dtac decided to do this project after just a one-hour talk with Chulalongkorn University’s Faculty of Architecture and Boonmee Lab,” said On-uma Rerkpattanapipat, Head of Communication & Sustainability at dtac. She is the key person behind the “Study Project on Mobility and Concentration of Thai Tourists during COVID-19 Outbreak based on Mobility Data” with the code name Project Phatthalung.
Ms. On-uma believes that the telecom industry is an important “data enabler” because only a few industries have access to such massive amounts of customer data. In the beginning, the three partners of the project thought about conducting the study in Phatthalung province only. But after reviewing the potential of mobility data, they agreed to conduct the study on a national scale.
Billions of Datasets
It took nearly two years for the Phatthalung Project to take off, though. The three partners spent considerable time assessing and reviewing the risks involved in mobility-data handling during the research process. Over a one-and-a-half-year period, they conducted Business Partner Due Diligence before signing a non-disclosure agreement, memorandum of understanding, and data processing agreement. At the same time, dtac wanted to make sure that customers’ privacy was respected by assessing potential data privacy risks and conducting an impact assessment.
Driven by a firm belief in the power of data, the project’s team members finally overcame all obstacles. The project got the green light on conditions that mobility data submitted for the study must be aggregated and anonymized with consent obtained from data owners. Importantly, the data was processed and analyzed in a closed, secured setting with no access to the internet. To do this, dtac dedicated a space at its head office to build what is called a “Co-Lab”.
Voravimon Srinut, Head of Customer Insights Department at dtac, has played a vital role in assuring that data submitted for the study were subject to the strictest data privacy policy. Call Detail Record (CDR) data was essential to the study, consisting of phone usage, time of usage, types of service used, and locations of phone usage. dtac anonymized the CDR data with One-way Hashing under a SHA-256 technique before submitting 10 billion data sets over a private network for the period covering June 2020-to-December 2021, which were classified to a subdistrict (tambon) level.
Enormous Datasets
Asst.Prof.Dr. Puripant Ruchikachorn, Co-Founder, and Rapee Suveeranont, Chief Technology Officer of Boonmee Lab, explained about the way of work that after receiving mobility data from dtac, Asst.Prof.Dr. Nattapong of Chulalongkorn University’s Faculty of Architecture would serve as the domain expert in the study. He was in charge of preparing a set of questions for data treatment. Boonmee Lab then reviewed the questions to determine if it was possible to provide answers with the mobility data in hand.
Mobility data from dtac naturally reflected behaviors based on two factors – time and location. Because such data are rather fluid, the study requires interpretations, literature reviews, and indicators in a bid to answer questions. For example, a tourist is defined by their movement from their province of residence or work to another province. Thus, the researchers need to set a time and space requirement to screen only trips that fall in line with the study’s definition. This ensures that the data used truly reflected the concentration of tourists and residents in a certain location. This required hard work and effective analysis from Boonmee Lab to deal with the enormous set of data.
Mobility data is considered “hierarchical”, showing the relationships between smaller and larger components in a database, which gives better understanding about the components of data and how they are related. Quantitatively speaking, there are many ways to interpret and categorize such data. For instance, it is possible to interpret data based on a day, a week, a month or even a year. On the location, these data can be categorized at a regional, provincial or subdistrict level. Moreover, it is possible to create urban- and rural-zone categories.
The research methodology used in the study was different from traditional ones, which mostly use surveys or interviews as a means to obtain data. Thanks to mobility data, the study was able to show tourist behaviors and movements in a certain area in real-time, allowing researchers to conduct research in various ways based on their topics. Mr. Rapee said at first glance, mobility data may look basic and uninteresting, but they are in fact very complex. The fun of conducting this study is to discover behavioral changes over different periods of time. Such discovery reflects the significance of data when interpreted in different dimensions.
“In this study, Boonmee Lab was flexible to an extent. We agreed to adjust some definitions based on variables or data presented during the course of study in our effort to get the answers,” Mr. Rapee explained.
Asst.Prof.Dr. Puripant added that Boonmee Lab did not know any personal information of dtac customers because the study was designed in a way that ensured his team only knew the beginning and the end of the mobility data of each anonymized customer. dtac randomly picked mobility data samples from about 10 percent of its customers for the study. Statistically speaking, such sample size is big enough to represent the population.
Hidden Potential of Mobility Data
Asst.Prof.Dr. Nattapong Punnoi, Urban and Regional Planning Department, Faculty of Architecture, Chulalongkorn University and a domain expert, added that the team members needed to work collaboratively with Boonmee Lab to understand data characteristics and consider the possibility of data analysis under dtac’s privacy policy. In the initial phase, the team developed a set of questions that could be analyzed by mobility data and may benefit tourism development. Subsequently, they analyzed structured data submitted by Boonmee Lab with Geographic Information System in order to visualize tourist concentrations, which is called a “scanning” process. This helped researchers understand the tourism environment and allowed them to discover movement patterns and tourist concentrations over time. Then, the researchers “connected the dots” of issues, leading to policy recommendations on tourism development in each province.
He stressed that the big challenge in this study was that the researchers must understand characteristics and the approach to use of mobility data, which then allowed them to set analytical framework and uncover possibilities and limitation of tourism development in each area.
Asst.Prof.Dr. Nattapong reiterated that mobility data has increased his understanding of the tourism environment with a detailed, accurate and comprehensive view – tourist concentration both overnight and day trip in certain areas and post-event evaluation in terms of the ability of attracting one-day and overnight tourists. In addition, mobility data also unlocks the creativity of the research team to set questions and analysis approaches to understand the tourism environment in each province, benefiting policymakers, investors, institutions and communities to design their tourism campaign, products and services more effectively.
“Thanks to the hidden potential of mobility data, three parties – dtac, Boonmee Lab and Chula – have decided to reveal the study of movement patterns and tourist concentration through the projected called ‘Tapping the Untapped’. This offers the other prospect of using data and broaden tourism opportunities in secondary provinces, boosting up local economies for the country’s sustainable development,” he added.
Co-Lab: Mobility Data for Social Good
Asst.Prof.Dr. Nuttapong stated that Co-Lab is a working model between dtac, Boonmee Lab and Chulalongkorn University in conducting a research and analysis of mobility data in order to better understand tourism circumstances for policymaking. This kind of tasks requires tremendous support and collaboration from a mobile operator, domain experts, government agencies and civil society to use mobility data in addressing social challenges and creating a better world. The constructive collaboration and mobility data are key to creating social innovation and driving sustainable development.
“The Co-Lab model can be applied to other social causes which mobility data allows us to get insights, including traffic congestion, healthcare, public space development, epidemic prevention and natural disaster management,” Ms. On-uma added. “This collaborative action on research and development will lead to a new dimension of policymaking, while the Co-Lab will provide an opportunity for experts and data enthusiasts to access and utilize mobility data to its fullest potential.”
She also hopes that this initiative will create big impact and eventually reshape the country at a structural level. The Co-Lab initiative, in all, resonates with dtac’s focus on sustainability and its ambition to create shared value with society at large and leverage its potential and strengths as a mobile operator for social good.