Periodic Reporting for period 4 - BigData4ATM (Passenger-centric Big Data Sources for Socio-economic and Behavioural Research in ATM)
Reporting period: 2017-11-09 to 2018-05-08
1. to integrate and analyse multiple sources of passenger-centric spatio-temporal data (mobile phone records, data from geolocation apps, credit card records, etc.) with the aim of eliciting passengers’ behavioural patterns;
2. to develop new theoretical models translating these behavioural patterns into relevant and actionable indicators for the planning and management of the ATM system;
3. to evaluate the potential applications of the new data sources, data analytics techniques and theoretical models through a number of case studies, including the development of passenger-centric door-to-door delay metrics, the improvement of air traffic forecasting models, the analysis of intra-airport passenger behaviour and its impact on ATM, and the assessment of the socio-economic impact of ATM disruptions.
• gathering and assessing the datasets to be analysed;
• conducting a literature review and a consultation process with ATM stakeholders, in order to define the research questions that would be addressed during the remaining of the project;
• developing methodologies and algorithms for the extraction of information about travel behaviour from non-conventional data sources;
• evaluating the applicability of the methods and algorithms developed through a set of case studies.
The work dealing data acquisition and quality assessment is documented in deliverable D2.1 Inventory and Quality Assessment of Data Sources for ATM Socioeconomic and Behavioural Studies. The main contents of D2.1 are:
• Data quality factsheets, which describe the characteristics of different data sources (quality, spatial and temporal resolution, etc.)
• A review of the data sources traditionally used for socio-economic and behavioural studies in ATM.
• A review of the different ‘Big Data’ sources that have been explored during the BigData4ATM project.
• A set of research questions to be investigated.
The data analysis and modelling work is documented in D3.1 Analysis of Passenger Behaviour from ICT-based Geolocation Data. The work includes the development of methodologies, algorithms and tools to:
• infer activity-travel patterns at different scales from mobile phone and Twitter data;
• extract intra-airport mobility indicators, such as heat and flow maps, from Twitter data;
• carry out opinion and sentiment analysis from Twitter data and extract indicators related to passenger satisfaction levels;
• extract expenditure patterns from credit card transactions data;
• validate the information extracted from non-conventional data through comparison with the information available from conventional, well-established data sources (e.g. surveys).
The methods and algorithms developed within the project were evaluated through four case studies, which are documented in D4.1 Case Studies:
• Passenger-centric door-to-door travel times. Results from this case study show that reliable estimations of airport catchment areas and travel times can be obtained. These results are useful to assess the integration of air transport with ground transport.
• Socioeconomic impact of ATM disruptions. Results from this case study show that novel ICT data can help to assess the economic impact produced by ATM disruptions, by identifying changes in expenditure patterns.
• Airports influence areas. This case study show the importance of access/egress in the competition between high speed rail and air transport. New data sources such as mobile phone records provide useful insights with respect to modal share and catchment areas at a lower cost than conventional methods.
• Air traffic forecasting. This case study explored competition between airports and how the opening of a high speed train connection can impact this competition, by cooperating with air transport and bringing new airports into the market.
• The design of more agile ATM system designs, that are more resilient to challenges such as rapid changes in demand or disruptive events, thanks to the ability to evaluate ATM performance through the impact on passengers.
• A seamless integration of ATM into the transport network.
• The use of new metrics and management decisions driven by passenger needs.
Despite being of an eminently exploratory nature, the research activities conducted by BigData4ATM entail a significant innovation potential and can open new market opportunities in several areas:
• New analytics products and services replacing or complementing the traditional methods used to gather information on passenger behaviour. These applications go beyond ATM, and can be of interest for all the stakeholders of the aviation sector.
• Innovative demand forecasting methodologies and tools. Here again, there is a clear applicability in the context of ATM, but also in other areas related to aviation business intelligence and more generally to socioeconomic research.
• New decision-support tools helping the ATM sector better respond to passenger needs and expectations. Particularly interesting are the opportunities to develop crisis management tools based on a more comprehensive knowledge of passengers’ reactions to ATM disruptions.