Data based planning optimizes public transport capacity utilization
- 1. DATA BASED PLANNING OPTIMIZES PUBLIC TRANSPORT CAPACITY UTILIZATION WITH FLOATING PHONE DATA ENHANCING CUSTOMERS’ TRAVEL EXPERIENCE Manfred Bock Global Industry Leader Rail - T-Systems International GmbH 1
- 2. SMART DATA: MANAGEMENT & VALUE CREATION Smart Data Management End-2-end behaviour Actual data from APTMS Individual Track construction work Intermodal IndividualizationAsset Data Base Status signals / crossings Passenger Information Journey planning Train operator Infrastruc- ture External/Public data: Weather / congestions / events / seasons Data Link 2 Predictive Analytics
- 3. SMART DATA SOURCE FLOATING PHONE DATA Information from cellular network ▪ Volume, frequency, dwell time ▪ Direction of movement, origin, destination, mode of transport ▪ “Overnight ZIP” Information from CRM-Database ▪ Age (age group) ▪ Gender ▪ Residence (ZIP Code) Mobile Data- sharing with Telco provider Geo Informa- tion Socio- demo- graphics Anony- miSation Transport PUBLIC services Retail Media & Advertising Deutsche Telekom CRM value- able Business Insights For Tourism ▪ Aggregated to the total population ▪ Conversion to „tile-grid“ 3
- 4. USE CASES TRANSPORT 1. COMMUTER FLOW ANALYSIS 4 Commuter Behavior & Traffic Measurement District 1 City center Industrial area 2Industrial area 1 Industrial area City center Subway/ railway LOCAL ▪ Commuter flow analysis (Origin/destinati on matrix) ▪ Route utilization ▪ Scenarios: simulating a diversion INNER-CITY ▪ Travel patterns ▪ Flow velocity ▪ Passenger transfer behavior ▪ Tourist flows ▪ District commuter volume ▪ Distribution of transport modes District 2
- 5. USE CASES TRANSPORT 2. ROUTE PLANNING Commuter Behavior & Traffic Measurement BENEFIT Visualization of volume of commuter flows: ▪ Capacity planning ▪ Flow velocity ▪ Effect of diversions City 1 City 4 City 2 City 3 Industrial area 1 Industrial area 2 City center District 2District 1 Subway/ railway Industrial area City center 5
- 6. FLOATING PHONE DATA SET – UP TRAFFIC MANAGEMENT SYSTEMS ▪ Design of transport management system ▪ Usage of empirical OD-Data from mobile networks ▪ Referred to 136 traffic districts in Karlsruhe Differentiation based on the day of the week Calibration of final destination ▪ Dynamization of demand Departure and arrival times referred to traffic districts Input variables for dynamic traffic assignment 6
- 7. PREDICTIVE ANALYTICS / FORECASTING MAIN COMPONENTS 7 Modeling ▪ Uses metrics that are already known ▪ Time of day, type of day, type of train, weather conditions, delays, etc. ▪ Modeling with non-self-learning processes: edge-node-specific histograms ▪ Modeling with self-learning artificial neural networks Creating forecasts ▪ Forecasting cumulative delays ▪ Use of probability density function ▪ Granularity specific to type of public transport (trains, buses, etc..) ▪ Deployment and evaluation of combinations of methods for forecasts ▪ Deployment and evaluation of methods in accordance with operational conditions
- 8. ▪ Travel itinerary is tailored individually and matched with personal preferences e.g. end-2-end travel plan, favoured means of transportation, etc. Personalized travel schedule Individual travel prognosis / Plan in real time ▪ Arrival and departure times ▪ Occupancy/capacity ▪ Real –time intermodal routing - Real-time recommendations for rout optimization Individual navigation Real-time travel optimization CREATING INDIVIDUAL PROGNOSIS/TRAVEL PLAN FOR COMMUTER ▪ Easy/not crowded access to stations and trains through better utilization of transportation 8
- 9. TARGET ARCHITECTURE historical data back-up HDFS models, forecasting HBASE preprocessor modelling forecast control forecast jobs KAFKA postprocessor Processed input data Train running forecast models forecast models forecasts result data of the forecast run Forecasts ForecastsPublic Transport Operator Train data Real Time Data Nighttime timetable planning data/dispositions data/master data (https file transfer) 9
- 10. BENEFITS AT A GLANCE FOR YOUR PASSENGERS AND YOUR OPERATIONS IMPROVED QUALITY OF SERVICE ▪ Travelers are informed in time – they can manage their multi-stage journeys them- selves ▪ High customer satisfaction GREATER COMMUTER FLEXIBILITY Real time information according to individual preferences SIGNIFICANT TIME SAVINGS Automation of travel information including delays forecast and movement forecasts saves time and improves quality of service DIVERSE USES FOR FORECASTS ▪ Dynamic customer information systems and mobile apps ▪ Integrated data for a range of multimodal travelling (rail, bus, air, car sharing…) ▪ Ensuring passengers make individual real time travel connections ▪ Optimization of multi-stage journeys ONE-STOP SOLUTION End-to-end offering – technological innovation, industry business solution, dynamic infrastructure – the key to digital transformation. The combination of public transport expertise and data analytics supports faster, easier decision-making. SMART FORECASTING METHODS Dynamic generation of real-time information, such as through integration of data on transport disruptions that affect the traveler’s journey 10
- 11. CONTACT Manfred Bock T-Systems International GmbH Global Industry Leader Rail Manfred.Bock@t-systems.com 11