Picture this: It’s 2010, and you’re in Paris, standing at the metro station waiting for a train to the airport.
Despite the morning rush hour, the platform looks deserted. Ten minutes go by — not the slightest clack of a train. Twenty minutes go by — your mobile phone emits a pathetic blip. The battery is down to 10%. No sign of a train, though.
“Excusez-moi,” you tentatively ask a loitering couple. “Do you happen to know when the airport train will arrive?” They look bewildered. “But it’s a public transport strike today, so *Parisian shrug* it could be any time, but not today.”
Back in those dark ages, Uber isn’t yet a thing. You need to rush all the way up the escalators, jump into the middle of the road to stop a passing taxi in its tracks, gesture deliriously that you need to get to the airport, and pray that there are no traffic jams on the way. Because you’re already late.
Today, your average traveler is far more connected thanks to the ubiquity of real-time big data in transportation. You know exactly when to leave, which mode of transportation to take, and what your contingency plan is if there’s a strike, a blizzard, an alien invasion, or any other type of traffic disruption.
Yet as cities grow denser and more crowded with private, public, and shared vehicles, managing the entire conundrum becomes not only harder but critical.
Time to shift gears: How will big data change the future of transportation?
According to the United Nations, there are 37 megacities today — dense metro areas with a population of 10+ million. By 2030, that figure is projected to increase to 47.
This highly concentrated form of urban dwelling we are shifting towards poses a host of challenges including resource supplies, waste management, and rising inequality.
But arguably the biggest issue (and at the same time, solution) is transportation management.
City congestion levels and population density
Source: Accenture — Society disrupted, now what?
In addition to people, world cities house a growing array of mobility players:
- Personal vehicles
- Public transport fleets
- Commercial fleets and last-mile delivery providers
- Micro-mobility actors (e-scooters, shared bikes, etc.)
- Mobility as a service providers
All of them navigate a city’s arteries at different capacities, with fluctuating demand, under varying weather conditions. Such ubiquity brings new dilemmas:
- What should we do about ride-hailing apps: ban them, regulate them, or leave them be?
- Does expanding shared bike infrastructure help decongest roads?
- How do we minimize the impact on existing traffic flows of constructing new infrastructure for shared mobility, e-vehicles, and public transport?
- And my personal favorite — Where should we put e-scooters?
Given the rapid growth in emerging mobility sectors such as ACES vehicles (autonomous, connected, electric, shared) and micro-mobility solutions, we cannot afford to leave those questions unanswered.
And this means we need to reconcile big data and transportation management.
There’s no shortage of big data in logistics, transportation, and urban planning. But that raw intel is rarely integrated into transportation planning activities, or is used only to a limited extent.
So how does big data affect the transportation industry at the moment? What can we do to put it to better use in the future? Buckle up and let’s go on a drive around the block.
Big data in transportation: Start your analytics engines with proven use cases
Transportation is complex because you need to orchestrate a cacophony of travel patterns into a coherent symphony of neat traffic flows — a task even Mozart would dread.
But the good news is that you get to use your instruments. And boy are there plenty:
- Smart city infrastructure
- Intelligent transportation systems
- Location-based services
- GIS/GPS data
- Predictive data analytics frameworks
- Machine learning and deep learning algorithms (AI)
Orchestrating a fine-sounding transport management scenario becomes a matter of selecting the optimal big data transportation use cases. We’ve got several lined up.
Transportation and logistics development
Cross-domain software solutions development for private, public, and commercial transportation
Big data analytics for transportation demand forecasting
If you’ve ever lived in a big city, you know the morning drill: leave at 7:30 to beat the traffic or take the 8:30 train if you overslept. If it rains, add an extra 15 to 20 minutes to your commute. If there’s a blizzard… well, maybe you should work from home.
Personally, each of us makes predictions about road traffic conditions. But professionally, why estimate traffic demand on a hunch when you can use historical big data for transportation planning and churn out a multitude of accurate predictions in a matter of minutes?
Here are the types of big data sources you can leverage for demand forecasting and transport planning:
- Data from mobile network operators (MNOs) such as call detail records (CDRs), floating cellular data, and shared smartphone location data
- Vehicle data from navigation apps (Waze or Google Maps), in-car navigation displays (TomTom), and fleet telematics
- Public transportation usage data from ticketing systems and smart cards
The best part? You don’t even need to use all of these transportation data analytics sources to get accurate predictions.
A group of Chinese researchers used call detail record data only to accurately map the physical travel patterns of residents in a city of 6 million. The team relied on a data fusion technique to form a labeled data set for supervised statistical learning. Then they used logic regression, artificial neural networks, and a support vector machine to create statistical classification models, producing on-target forecasts.
Traffic congestion management
Traffic jams are the bane of many city dwellers and a locus of profound pressure for city managers.
Road premiums, public transport discounts, narrower one-way lanes, tunnels, “naked” streets without traffic lights, even flying cars and drone-based logistics — there’s no shortage of realistic and aspirational ideas for combatting traffic congestion. Some work to an extent. Others flop. Yet no city is completely free of traffic jams.
How come? Because no two urban layouts are alike. Traffic congestion may bear painful similarities across the globe, but the root causes differ.
Some locations are more prone to congestion-inducing weather conditions — rain, fog, snow, hail, and other natural occurrences are the main culprit of an estimated 15% of traffic congestion cases in the US.
Suboptimal physical infrastructure design leads to recurring congestion and continuous bottlenecks at popular locations. But reducing the hourly throughput at such locales is either not feasible or too costly to implement.
While it can’t fix physical causes, you can optimize the digital end of this conundrum with a big data-fueled intelligent transportation system or more targeted traffic optimization solutionsБ:
- Locate inefficient parking layouts resulting in bottlenecks
- Devise better routes for commercial truck passage and urban deliveries
- Improve traffic light signal timing
- Optimize micro-mobility and shared mobility solutions
- Entice drivers to ditch cars with effective multi-modal transportation routes
Technologically, all of the above use cases of big data in transportation and traffic engineering are already feasible.
Did you know that we’re wired to tolerate a one-hour commute per day maximum (roughly 30 minutes each way)?
Better known as Marchetti’s Constant, this 30-minute travel time has been shaping city layouts and dwelling patterns for centuries (and our tolerance for paying exorbitant rent if it means a shorter commute).
However, rising urban density and subsequent traffic congestion undermines our evolutionary inclination to stay within the 30-minute mark. Some cities choose to double down on public transport network development and impose various restrictions to make personal car use an expensive choice, which can be a viable long-term solution.
But commute times need fixing today. A combination of big data and analytics for intelligent transportation systems can provide immediate relief. By operationalizing available sources, you can:
- Develop more accurate traveler systems and journey planning apps
- Design policies to incentivize public transport use rather than penalize citizens
- Improve city planning and expand road infrastructure to accommodate known commuting patterns
- Integrate mobility as a service and shared transport projects into the city’s public transport repertoire to diversify commuting routes
Check out a detailed case study about building an IoT-powered mobility as a service solution
Singapore, a city consistently praised for the best transport system, designed a transportation route layout where 90% of the population lives within 300 meters of a bus stop. Next, they effectively integrated the bus network with the city’s metro and light rail systems to promote multi-modal journeys. These accounted for 67% of all journeys as of 2018.
Now, Singaporean officials are tackling road traffic management. In 2017, they rolled out a transportation data analysis system capable of aggregating and analyzing real-time road traffic data using Global Navigation Satellite System (GNSS) technology and in-vehicle telematics data.
By mid-2023, the city-state plans to switch to a satellite-based electronic road pricing (ERP) system for collecting road tolls and dynamically managing road prices. An onboard vehicle unit will provide drivers with real-time information on ERP charging locations and rates, overall traffic conditions, and traffic zones with special speed restrictions.
Intellias too has been working on the future of commute optimization as a technology partner for an automotive company. We’ve been helping our client establish low-latency processing of geospatial data and deploy custom geolocation solutions to the market that provide urban planners with precise operational insights.
Most every driver has received a parking fine on the windshield. But the truth is that parking tickets are not the real source of frustration — lack of available spots is.
Insufficient parking also generates extra costs for municipalities:
- Helpless circling around a 15-block area in Los Angeles alone produces over 730 tons of CO2 a year and leads to an extra 47,000 gallons in fuel consumption.
- Researchers also found that cruising cars negatively affect overall travel flows. Decreasing the proportion of cruising vehicles from 50% to 0% could decrease total travel times by 85%.
On the other hand, overparking also has negative implications on housing availability, plus it reinforces car dependence.
That’s why the parking equation requires careful balancing that data science in transportation could help accomplish:
- Parking demand prediction. MNO and free-floating cellular data can be used to determine where drivers search for parking and cruise the most.
- Parking management. Smart parking apps can transmit data to drivers and connected cars to facilitate navigation to the closest available spot and process payments.
- Parking violations. Collected parking data can be used to identify areas with high violation rates and look into improving the situation.
For instance, our team has been working on a smart parking solution for one of our clients. The proof of concept mobile app navigates drivers to the closest free or paid parking spot and builds a further walking route to the selected destination. Convenient for drivers? Yes. Valuable for urban managers? Undeniably.
Bike and pedestrian infrastructure development
Four wheels or two wheels? People are now choosing the latter.
Heightened anxiety over public transportation and a surge in exercise has meant that more and more are choosing to use one of the most basic forms of mobility, leading to a so-called ‘bike boom’.
From Sydney to San Francisco, urban dwellers are hopping on bikes to get around town. That’s not just a healthy trend but also an opportunity for city managers to offload the roads.
Many governments are incentivizing this habit to make it stick. Italy and France have established stipend programs for bike purchases during the pandemic. British city planners are rolling out pop-up cycling infrastructure and temporary lanes to promote greater bike use.
But to make the biking habit stick, even more cycle-only corridors will have to be incorporated into urban layouts. Once again, different big data and machine learning transportation models can be used to create efficient infrastructure for riders and walkers alike:
- Cycling route visualizations. To locate good locations for new cycling lanes, planners will need to identify common travel patterns. Geospatial data visualizations and analysis are a strong contender for this. For example, StreetLight Data recently visualized the most-biked streets in San Francisco using city sensor data. By infusing their big data transportation system with biking and pedestrian tracking, the team hopes to reduce the number of collisions, casualties, and fatalities, plus promote the development of better cycling infrastructure in the region.
- Transportation planning and infrastructure investment analysis. To develop comprehensive cycling plans for cities, managers require more insights into residents’ walking and biking activity and short-distance trips. Current applications of big data in urban transportation already allow for obtaining granular insights to locate underserviced areas or unsafe environments and modeling various infrastructure cost scenarios.
- Car-free and pedestrian zone planning. 3D models and geospatial data analysis can help establish car-free zones without disrupting car travel. For instance, data from Google Street View can be used to count pedestrians and can be augmented by logistics big data to determine the costs of converting a street into a pedestrian zone.
Case in point: Swiss Federal Railways leveraged a combination of CCTV video data and IoT data to analyze pedestrian flows out and about the Lausanne train station. Using 100 sensors, the planners mapped common trajectories, behaviors, and flow dynamics. Then they used anonymized data to plan future pedestrian infrastructure.
Why is big data in transportation not mainstream yet?
There are a lot of technologically feasible ways to use big data in logistics, transportation, and urban planning, as the above use cases illustrate.
But last time I checked, parking a car still was an issue in highly crowded areas. So what’s the deal?
Extracting and processing transportation big data is complex. Most cities don’t have sufficient road sensors to track traffic flows. Capturing and mining video data requires a robust, low-latency data governance and management platform. Fusing multi-source data, obtained in different formats, also requires technical infrastructure maturity not all organizations possess.
Then there’s a far more important question of personal data privacy. No one wants to get caught on tape and then recognized by their data science colleagues as that guy struggling to park on the lot. All transportation big data should be cleansed and anonymized prior to use, then protected with top security methods. After all, a data lake, containing records about an entire population’s daily movements, is a lucrative tool for misuse and a target for hacking.
But the better news is that both issues are addressable and don’t detract from the magnitude of positive changes wider adoption of big data in transportation could create for us individually and as a society at large. Shall we build a better ride together?
Contact Intellias to learn more about adopting and using big data in the transportation industry.