Further, Faster, Cheaper: How Data Analytics Helps Mobility Businesses Go an Extra Mile
DOWNLOAD THE GUIDE
DOWNLOAD THE GUIDE
DOWNLOAD THE GUIDE
When someone drops the term ‘data science,’ trains aren’t probably the instant association. But, as the story of Herman Hollerith demonstrates, data and computing sciences have been tied to mobility for centuries, as they continue to do today.
We must move back to the 1880s USA to understand that curious linkage. At that time, the country’s population was crashing and soaring, stimulated by economic development and millions of immigrants pouring in. Keeping track of the demographics was a daunting but important task that relied mostly on the data from national census counts held every ten years.
The counts were collected by pen and paper, which the US Census Bureau was desperate to change. It needed a new solution: the 1880 census dragged on and would take seven years of arduous work. Given the rapidly growing population, the 1890 count would last even longer.
Hollerith, a young talented engineering graduate, was charged with the task of coming up with a more efficient alternative for recording vast data volumes.
From a train ride to data revolution…
As with many great inventions, the source of his inspiration was unlikely. It lay in… public transport.
Noticing how conductors punched tickets for his fellow passengers spurred Hollerith to combine perforators with punch card reading systems. They were already used in advanced looms to enable easy collection and storage of large volumes of data.
After some years of experimenting and refining, he produced what would become a predecessor of modern computers: the Hollerith tabulating machine. The invention passed its first major test with flying colors: thanks to the tabulator, the 1890 census was completed in just three months, covering additional details, such as the number of children in each family.
‘All right, but what does this have to do with IBM?’ you might ask. Here’s the answer.
Encouraged by the success of his invention, Hollerith founded his own business, Computing, Tabulating & Recording Company. After a series of mergers, the company would turn into a household name: International Business Machines Corporation or IBM. And all that happened partly thanks to that one moment of inspiration on a train journey.
…from data to mobility innovation
Hollerith’s story demonstrates how innovation in using data can lead to new, exciting opportunities, which brings us back to today. The world, where 59.5% of organizations use data to drive innovation.
With that, data science and analytics are paying back the coincidental “debt” they owe to transportation, as organizations around the world use the information to reduce costs and the environmental impact of mobility while increasing its safety. Understanding why data science matters and how it can benefit the automotive industry is paramount to the mobility companies’ success and central to this publication.
How data science and analytics power modern mobility
In this piece, we explore how and why data science and analytics are fundamentals of future mobility. We will provide essential terms and definitions, explain how mobility companies can benefit from using data effectively, and supply specific, real-world use cases documenting the endless opportunities in data management and analytics for mobility companies.
Our guide is divided into three main parts. The first will explain the basic differences between data analytics and data science. Then, we will briefly discuss the many benefits of outsourcing data science and data analytics. Finally, we’ll serve you specific mobility data science and analytics use cases to demonstrate how you can apply both fields to improve profitability and retain customers.
Anyone who’s ever tried to study the subject of data will quickly stumble upon two deceptively similar-sounding terms: data science and data analytics. Even though both concepts have been around at least since the 1960s, for many, they’re still a source of confusion. So, what are they, and how do they differ?
Data analytics involves studying large chunks of information, searching for patterns and trends to make predictions and solve specific, defined business challenges. Another important aspect of data analytics is visualization—presenting complex findings so that it’s understandable and easy to work with for non-technical business people and decision-makers.
Data science deals with gathering, integrating, modifying, and manipulating data. Its primary goal is to develop statistical and predictive models that help organizations ask new questions and learn what they don’t know. To that end, data science uses artificial intelligence (AI), machine learning (ML), statistics, advanced analytics, mathematics, and programming.
Most data science projects comprise six key stages:
Defining a problem
Data science usually deals with the unknown, which leads to questions appearing during a project. That said, preliminary knowledge of the problem is essential to set the project’s scope, methods used, and resources needed and to develop metrics to measure the effectiveness.
Gathering data, both structured (organized and stored in a set format, e.g., spreadsheets, lists of names, addresses, phone numbers) and unstructured (without one predetermined format and usually harder to process, e.g., emails, images, handwritten notes, audio files).
No matter how well-tuned the model is, bad data will lead to bad results. That’s why the data must first be cleansed (e.g., by removing duplicate records, standardizing measures and null values, or deleting obsolete entries), aggregated, and structured. This stage is often the longest and most cumbersome but, simultaneously, crucial for building a reliable data model.
Developing predictive AI/ML models for an in-depth analysis of the collected datasets.
The models are assessed and compared to find the most consistent and precise. Depending on the results, the model can be tweaked and tuned to improve performance or enable it to work with previously unseen data.
Deployment and maintenance
Once the model is developed and evaluated, it can be applied to provide insights, find problems to solve, and otherwise generate value for the business. Its performance should still be monitored to implement adjustments when necessary.
More recent telematics solutions are increasingly focusing on mitigating or eliminating human error. They provide in-depth insights into driving patterns, collecting dozens of parameters to recreate a particular driver’s behavior given specific weather and traffic conditions, time, and topology.
The current AI-based telematics solutions still use the ‘traditional’ ABC metrics. However, they supplement them with driver behavior analytics (complying or violating traffic laws, taking a break when tired, speeding offenses, etc.), augmented with the context of driving (road conditions, traffic, weather). Additionally, they calculate the total risk by considering the frequency and magnitude of risky events. Lastly, the final risk score is provided, based on which insurers seeking market differentiation can build their innovative products and services.
Although data analytics and data science both deal with data, statistics, and predictions, they differ in how they approach information and what tools they use.
Below, we provide a list of the most crucial differences between both. For a more in-depth analysis, check out our data analytics and data science comparison.
Data analysts are presented with defined problems to solve and work reactively to find answers using data. On the other hand, data scientists work on datasets proactively to discover new issues, questions, and approaches to using data to solve business challenges.
For example, a data analyst may be asked to investigate sales data to find out why one region performs better than another, why the last marketing campaign was unsuccessful, or how to improve current production processes. Since data analysts can specialize in various fields of business activity, they carry different job titles: financial analyst, sales analyst, marketing analyst, operations analyst, and more.
Data scientists ask new questions throughout the project and find issues no one even suspected existed. Then, they create tools, such as AI algorithms and predictive models, to discover insights and solutions to the newly found challenges.
The type of data both fields work with is directly related to the questions they aim to answer.
Data analysts know the problem from the outset, and their organization has already collected and organized the dataset needed to solve it. The data is structured and well-defined, just like the challenge itself.
For data scientists, the process is more like exploration, while the datasets tend to be larger and consist of both structured and unstructured records. For this reason, an essential part of a data scientist’s job is to develop tools that will enable them to manipulate that data or make the process easier.
Different tasks call for different abilities. Even so, there’s a significant overlap between the skillsets of data analysts and scientists, and proficiency in all abilities listed below is necessary regardless of the role.
Unsurprisingly, data analysts are more focused on the analytical stages of data projects, which makes mathematical and statistical experience required for their job. Business knowledge is just as important, especially in the area relevant to the specific task. Other skills that will come in handy include data warehousing, data mining, and visualization tools expertise.
When compared to data analytics, data science has much more in common with software development. The knowledge of coding, computer science, automation, AI/ML modeling, and programming languages is crucial. Then, data scientists must also be able to apply these skills in the context of the organization they work at.
Mathematicians and statisticians first formulated the concept of data science in the 1960s. But it wasn’t until the 2010s that their findings could be leveraged commercially on a broader scale to inform decision-making and help find solutions to business challenges. Tapping into decades of research became possible thanks to the popularization of the Internet, digitalization, increasing computing power, and cloud-based solutions.
Nowadays, we produce 2.3 trillion gigabytes of data daily, which presents a rich source of information for enterprises seeking to find answers to business challenges. Also, organizations have already been collecting and storing data for years in one form or another—be it invoices, emails, or reports. All this means that businesses from virtually all industries now have more opportunities than ever to use all that data to their advantage.
Unknown, usually discovered during the project
Structured, finite datasets
Structured and unstructured, larger volumes
Mathematics, analytics, business intelligence tools
Programming, AI/ML, computer science
As our ability to collect and analyze data increases, businesses from all sectors can tap into data to improve their operations:
Banking and finance
Financial institutions have long had access to large volumes of customer data. Today, banks such as HSBC use AI to tap into that data, learn more about their clients, microsegment them, and create custom financial products. Other use cases include risk management, developing services for traditionally unprofitable clients, supporting investment decisions, improving customer experience, and enabling blockchain-based solutions.
Obtaining data is more challenging in agriculture and farming, but research to develop new collection methods has been going on for years. Companies like Bayer utilize drones, sensors, and satellite imagery to gather data that AI models then turn into real-time insights. Farmers in Taiwan, for instance, use digital farming methods to overcome climate change. Big-scale data collection projects are also underway in some of the world’s poorest countries to aid farmers and fight hunger. Some other use cases are precision farming techniques, sustainability initiatives, weather and climate predictions, optimizing equipment and resource usage, and supply chain management.
Healthcare and biotechnology
Modern medicine proves best that data science can literally save lives. Biotech concerns like Astra Zeneca have entire departments of data scientists devoted to accelerating clinical trials, predicting molecular reaction outcomes, and improving our understanding of diseases. Data science can also be applied to genetic data processing, outbreak prediction and containment, targeted medicine development, drug research, and logistics management in clinics.
Public services and security
Governmental institutions invest in data science initiatives to inform their decisions. Companies like DataKind specialize in working with public agencies, helping them improve access to sanitation, redevelop underserved areas, and empower marginalized members of society. Other examples include transportation and infrastructure management, law enforcement, data security, and military intelligence.
In the energy sector, data science can ensure supply and distribution stability, e.g., by AI-powered demand forecasting or grid loss predictions. Data science also assists energy enterprises through production and performance monitoring, geographical data insights, efficiency predictions, cost-effectiveness analysis, weather predictions, and renewable transition assessment.
Data science is utilized throughout the manufacturing process, from sourcing raw materials to shipping the final product. Unilever, for example, uses data to enhance supply chain management, production planning, and machine control. Data solutions can also be applied to optimize maintenance, produce demand forecasts, and improve customer service.
Telecom companies play an important role in developing data technologies but also use these innovations. AT&T, for example, collaborates with Google, Microsoft, and Amazon as research partners. Simultaneously, the company uses data-driven solutions to prevent network outages and fraud, block robocalls, and offer self-service functionalities. In telecom, data science can also help prevent customer churn, improve user experience, and develop new product offerings.
Marketing, sales, and advertising
Knowing your customer is key in sales and marketing, and data science can help companies enhance that knowledge. A global advertising firm Wunderman Thompson teamed up with IBM to use data science for a more effective customer insight discovery and increase their client’s ROI. Other use cases of data science in marketing and sales include conversion analysis, marketing campaign predictions, and evaluating the effectiveness of sales and marketing efforts.
Even small details can make the difference between customers choosing a store over the competition. Aware of that, chains like Walmart analyze every factor affecting their sales—such as major events, social media trends, shopping patterns, and weather—to optimize their offering. Data science also allows retail businesses to improve customer segmentation and store planning, develop targeted loyalty programs, and optimize prices.
When it comes to the hospitality sector, the main role of data science is improving operations and discovering new ways to delight customers. As an industry leader, the Marriott hotel chain excels at using data for both with its data-driven dynamic pricing, targeted marketing, and loyalty programs. The use cases for data science in hospitality don’t end there—some extra examples are customized discounts and offerings, guest interest forecasts, and optimized accommodation.
Data science holds much potential, and more and more companies are beginning to realize its benefits, a fact reflected in the growth of the data science market.
With that in mind, you may be tempted to start looking for a data specialist for your company or maybe even to think about recruiting an entire in-house data team. But while data science is a promising investment, there are a couple of caveats that may make you reconsider your hiring plans.
According to Glassdoor data, the average salary of a data scientist is USD 125,724 in the US or EUR 70,000 in Germany. This significant sum doesn’t include bonuses and other costs companies must cover to maintain an in-house position. Moreover, these costs will only add up if you decide to scale up your project by hiring additional data scientists.
Hiring and training new team members cost money and time. Even if you find an experienced data scientist, introducing them to the specifics of your business and industry can take weeks.
Then there’s also the time your newly-hired specialists will need to research your data sets and develop custom tools. Combined, all these roadblocks can significantly delay the time before your data project takes effect.
Data science and analytics involve various skills and business knowledge. Since the data science boom has begun relatively recently, the demand still surpasses the number of experienced data professionals. Due to that, finding data scientists and analysts who will live up to your expectations may not be that easy.
The demand for data science services tends to be very dynamic, even within the boundaries of a single organization. This means that at one moment, your data experts may be up to their ears in major predictive model development, only to be left with very little to do once the project is complete. Due to that fluctuating demand, full-time in-house data teams may not be sustainable in the long term.
Lack Of Experienced Specialists
All the above challenges may lead businesses to postpone the launch of their data science projects or even give up on them completely. However, data science outsourcing is a viable alternative to building and sustaining an in-house data team.
External companies that provide data science as a service (DSaaS) solutions have everything necessary to perform all data-related tasks, including data collection and cleansing, modeling, automation, analysis, and more. They offer a range of benefits for organizations that can’t afford or don’t want to risk hiring their own data scientists.
With the average salary range of EUR 70,000 (in Germany) to over USD 125,000 (in the US), hiring a full-time data scientist may be challenging, especially for smaller companies or those uncertain about the scope of their data projects. But matching recruitment with the time and effort required to complete data science tasks requires experience. Otherwise, you may get too many specialists who will end up on (a costly) bench or hire too few, slowing down the progress.
Instead, DSaaS outsourcing partners work on a per-project basis. This means that launching a one-time project won’t generate additional costs after work is complete. Additionally, an outsourced team can work dynamically, allowing you to add new specialists or redirect the current ones to scale up or down as needed. Cooperation with an established outsourcing partner allows you to skip the lengthy recruitment process if you need additional manpower when the project is already underway.
And don’t forget that as the demand for data scientists increases, finding a specialist with adequate experience can be a challenge in the first place. And even if you manage to hire your data scientists, the highly competitive market can make it difficult to retain them. Losing a specialist you spent lots of time finding and training is bad enough; it can get even worse if it happens during a key stage of your data project.
On the other hand, DSaaS companies warrant reliable access to adept data experts because they only focus on data science and analytics. To stay competitive, they continue to engage the best data scientists and analysts out there, so you can be sure that your project will be in good hands from day zero. Additionally, third-party data specialists don’t have to spend weeks on training or upskilling; after a short onboarding, they can apply their expertise without burning your money.
This focus also compels DSaaS providers to innovate and stay up-to-date with the latest developments in data science. Doing that internally is costly and difficult due to the scarcity of specialists and the complexity of the domain. For many smaller companies, keeping abreast of data science innovation would require refocusing resources away from their main business activity. But with a reliable outsourcing partner on deck, you gain access to new findings in data science that can grant you an edge over competitors in your industry.
Outsourcing gives companies of all sizes an opportunity to utilize their data. Still, there are three crucial things you need to talk over with your DSaaS partner before signing a contract:
IBM estimates that the global average cost of a data breach is USD 4.35M. To prevent costly leaks in a situation where loads of potentially sensitive data are passed from hand to hand, ensuring anonymity, compliance, and security is essential. Make sure to ask your outsourcing partner how they plan to address all three aspects of data protection and work together to establish necessary protocols and safety policies.
Ensuring data security will only be possible once you know who manages particular data sets and share that knowledge with your data science provider. Actually, a transparent data ownership structure is crucial for any data science project for many reasons, such as accountability or maintaining adequate data quality. A good way to start is by defining data domains (i.e., the data sets that are most important to your organization) and then outlining ownership over each domain.
Efficient communication is the foundation of any initiative. Specify what channels you’ll use, when all relevant stakeholders can be reached, and what rules to follow when discussing the details of your data science project. A clear communication ruleset will warrant everyone is on the same page, both in your organization and on the side of your external partner.
By this point, you have a basic understanding of data science and analytics: their definitions, the crucial stages of any data projects, and why delegating them to third-party data science technology experts may work best.
We’ve also listed several data science use cases from various industries. Now it’s time to take a look at data science from the perspective of the mobility sector. What benefits does it offer to mobility businesses? How can the automotive industry apply automotive data analytics in practice?
Let’s find out.
If you were to ask any of your colleagues to name one practical application of big data, there’s a good chance that their answer would be “predictions.” And they wouldn’t be wrong—making accurate forecasts based on vast sets of records is one of the primary uses of data science.
Knowing how likely an event is to happen in set circumstances gives businesses a chance to mitigate risks by recognizing threats before they even occur. In banking, predictive analytics is used for fraud prevention; in supply chain management, it helps foresee shortages; and in agriculture, temperature forecasts allow farmers to protect crops from adverse weather. But what about the mobility industry?
Any commercial fleet manager will tell you there’s no bigger roadblock than a vehicle suddenly breaking down in the middle of nowhere due to an unexpected failure. Each such event has a range of consequences: the vehicle has to be towed away and repaired, its lifespan and residual value get shorter, a replacement vehicle needs to be dispatched, and the associated costs escalate due to vehicle downtime. Worst of all, every mile a vehicle goes with an undetected fault further increases the likelihood of a dangerous accident.
Of course, fleet operators try to prevent all that by running regular servicing. This system does work to some extent but is far from flawless.
In 2022, US fleets reported a 10% increase in maintenance and repair costs. As this upward trend continues, managers need a more proactive and efficient approach than scheduled maintenance to keep their fleets up, running, and profitable. Data science as a service platforms provide a solution.
In the traditional approach, vehicles undergo maintenance checks either periodically (e.g., once a year) or based on mileage (e.g., every 5,000 miles). Crucial contextual factors, such as driving style or road conditions, are not accounted for in that approach, even though they have a major impact on component wear and contribute to the likelihood of breakdowns—and their severity.
Besides, relying on scheduled repairs can lead to unnecessary downtime when a perfectly fine car spends time in the repair shop rather than on the road. All in all, time- and mileage-based maintenance scheduling can negatively affect the profitability of a commercial fleet.
Automotive analytics helps fleets address both of these issues. For that purpose, it uses diverse data such as GPS readings, CAN bus data (Controller Area Network, internal system car components used to communicate with each other), and maintenance records.
On-board IoT sensors are important for predictive analytics, feeding fleet analytics software with real-time vehicle data. With the aid of these devices, AI models can factor in fuel and oil usage, track the condition of every connected part, and receive notifications about mechanical failures through Diagnostic Trouble Codes (DTCs).
Another essential source of information is mileage relative to road conditions, i.e., the distance covered in specific conditions that may affect the health of car components like suspension or tires. Finally, there’s driver behavior, such as harsh braking, which can also put additional strain on the car.
Scheduling based on
Time since the last maintenance, mileage
Dynamic factors (component wear, road conditions, driving style, etc.)
Prevents breakdowns and helps keep the vehicle in good condition (but can result in unnecessary downtime)
Increases safety, prevents breakdowns, improves vehicle uptime and lifespan, helps preserve residual value, generates savings
Has no setup costs but generates higher operational costs and can negatively affect residual value
Brings long-term savings after the initial investment in telematics devices and data analytics implementation
Predictive analytics fleet management algorithms process a combination of vehicle, driver, and external data to suggest improvements, allowing businesses from the mobility sector to:
Considering the above benefits, it’s easy to see how predictive vehicle maintenance is a perfect solution for any enterprise whose daily operations depend on the smooth functioning of large fleets: delivery services, public transportation organizations, car rental and car-sharing companies, public services, or long-haul truck operators.
We’ve mentioned how predictive car maintenance can go a long way to keep cars in use for longer. And although there’s no way to stop aging completely, automotive data analytics gives you a couple of tools to slow down this process significantly—and with it, the rate at which the residual value of vehicles degrades.
If you’re unfamiliar with the term, the residual value represents the vehicle’s value at the end of the lease period or at the time of sale to its original price.
It describes how much the vehicle has depreciated over time; if you lease a car for three years, its residual value will reflect its worth after three years.
As automotive technology progresses and new manufacturers enter the market, metrics such as mileage, car age, and maintenance are no longer enough to estimate a vehicle’s residual value accurately. Fuel consumption, brand, and market trends are just some factors that need to be considered. Contextual aspects that contribute to vehicle depreciation, such as driving style or road conditions, while extremely impactful, are often neglected.
The good news is that fleet data analytics systems can process these factors to provide insights on minimizing the decrease in residual value.
To understand how to protect your vehicle’s residual value, it’s useful to think of a car as a sum of its components. In the most extreme cases, a single part can account for as much as half of the total price of a used vehicle.
Telematic devices help track the wear and tear of individual components, but that alone won’t optimize the residual value.
Instead, fleet management analytics finds patterns and relationships between sensor data and contextual factors such as weather, road conditions, or driving style. Based on that, the algorithm evaluates how each affects every single part and, effectively, the car’s overall health. Here are some examples of how driver behavior and external conditions can impact specific car parts:
Hilly terrain and often harsh braking
Brake pads overheating, brake disc wear
Low-quality roads and excessive speed
Suspension damage, tire wear
Brake wear, body bumps
Knowing these and other ways in which various contextual factors impact the car’s value is half of the success towards reducing its depreciation. The other half are the actions undertaken to optimize routing, improve maintenance, and coach drivers. Altogether, they can improve each vehicle’s performance and minimize repairs throughout its lifespan to slow down its depreciation and protect the residual value.
Transportation is responsible for around 16% of greenhouse gas emissions globally and 27% in the US. In the context of climate goals set by the UN—to cut them to net zero by 2050, mobility businesses are under heavy pressure to reduce their footprint.
The mobility industry has already made some steps toward that end. For example, between 1995 and 2015, car makers managed to reduce the CO2 emissions in new cars by 36%, from 186g/km to 119.6g/km. Fuel efficiency is also improving, with 10 out of 14 top car manufacturers in the US improving the Mile per Galon (MPG) metric in their new models.
At the same time, recent findings revealed that car makers underreport half of the emissions declared on average, which suggests that the mobility industry may need to innovate beyond just reducing fuel consumption and carbon emissions mechanically if the climate goals are to be achieved.
One of the tools mobility companies can use to bring down their footprint is automotive data analytics. There are several ways in which it can help them protect the planet, meet environmental targets, and optimize operations. Since all of them essentially come down to protecting the environment by reducing and optimizing the use of fleet resources, it’s easy to see how they help any mobility business become greener and more profitable.
Picking the right itinerary can make a huge difference in how much fuel a vehicle will consume and, by extension, what effect it will have on the environment. Here, predictive models consider factors such as traffic, road quality, infrastructure, and weather conditions, to come up with the best route in terms of eco-efficiency.
There’s a limit to how much load every vehicle can carry, but AI-powered car tracking systems help identify opportunities to transport more cargo with fewer trips. For instance, they can connect to online freight exchange platforms and show how to utilize an empty trailer on the return trip from a delivery drop-off. And where’s the emissions-saving potential? Quite simply, fewer trips, especially on an empty truck, translate to a lower carbon footprint.
Similarly, data-based fleet analytics solutions can contribute to environmental protection by helping to match the size of the vehicle to the load it carries. However, matching cargo to a fitting car can get tricky as the fleet grows in size and diversity. In the long run, this can harm fuel consumption.
Mobility analytics systems can review your past journey records and create the optimal car-route-load combination. AI models can also assess the viability of downsizing vehicles for particular trips or switching to EVs.
Sometimes, vehicles use too much fuel for purely technical reasons. For example, underinflated tires have lower rolling resistance, which leads the car to consume more fuel. Another cause may lie in improper engine oil. Switching to a lower-viscosity lubricant has been proven to reduce fuel consumption by up to 1.5.% for a class 8 truck. Over a year, such a reduction would save hundreds of dollars per vehicle and significantly reduce fleet footprint.
How can an automotive data analytics solution help here? For instance, by providing fleet managers with efficiency reports and identifying potential upgrades. It can also be used to test new components’ performance in fuel economy, creating opportunities for further improvements.
Faulty pumps, dirty fuel injectors, and other fuel system failures may increase the fuel consumption of a single vehicle by just a little, which makes them easy to overlook. Things get worse if the problem occurs regularly or on a fleet scale. Fleet analytics software can easily prevent that by tracking part conditions and suggesting the optimal servicing schedule.
Driving style is a crucial factor that can contribute to fuel overconsumption. Excessive speeding, rapid acceleration, harsh braking, and other aggressive driving behaviors can all be linked with increased fuel usage. The same is true for idling, which, according to the US Department of Energy, uses between a quarter and half a gallon of gas (roughly 1-2 liters) per hour.
With the help of fleet data analytics, you can predict and track the costs of the unsustainable driving style. Next, you can implement driver coaching programs with automated, real-time feedback alerts sent to drivers when they engage in aggressive maneuvers or gamification elements and incentives to encourage and reward eco-friendly driving.
Six Steps To Promote Eco-Driving
The shift away from internal-combustion engines (ICEs) is already happening worldwide. EV sales continue to surge in the US, Europe, and—particularly—China. More and more countries and states are announcing legislation to limit or ban the use of ICEs. At the same time, they’re coming up with new ways to encourage drivers to switch to EVs, from exclusive parking tariffs and zones to subsidies.
Car manufacturers, aware of customer trends and upcoming legal deadlines, are also taking steps. General Motors, for instance, has pledged that by 2035, it will phase out the production of all non-electric vehicles, and by 2040, it’ll be wholly carbon neutral. Other major producers, like Ford, Honda, Toyota, and Volkswagen, aim to achieve carbon neutrality by 2050.
The pro-EV movement hasn’t gotten past mobility businesses who, just like private car owners, have all the reasons to go electric. Governments and municipalities offer a range of incentives for companies that make the transition. At the same time, the cars perform well and require less maintenance; they’re sustainable and consume less fuel.
But in the case of enterprises, the switch isn’t as easy as replacing one vehicle. There are three main concerns stopping fleet owners from adopting EVs.
Why Do mobility businesses fear the transition to EVs?
Total Cost Of Ownership
The EV transition requires a steep initial investment in upgrading the fleet and the infrastructure to support it. However, these expenses can be partially covered by subsidies; in the long run,savings will more than make them up.
Battery capacity is another important consideration, but the average EV range steadily grows yearly. Planning and driving style adjustments can extend the range even further.
Power station scarcity and charging times require fleet managers to optimize routes and timetables.
A successful transition to electric fleets requires elaborate, long-term planning—and automotive data analytics solutions are perfect for informing those long-haul plans. To do so, they use information collected by telematic devices and external apps, systems, and equipment, among others:
Knowing where and for how long a car is out of use allows the algorithm to sync that time with charging. For instance, drivers can take breaks during trips as the battery charges.
Tracking and analytics enable fleet managers to see how the individual driving habits of each driver will affect the range of your EVs. Based on that knowledge, they can introduce coaching programs to improve driver performance.
The average mileage of vehicles determines the range of EVs a fleet will need. It’ll also allow for optimal route planning.
If necessary, vehicles can use charging stations available along their itinerary. This will also tell you what upgrades your grid and infrastructure will need.
Based on that and other data, fleet analytics software can provide a range of insights that will inform various aspects of your EV transition, such as:
A custom combination of EVs, ICEs, and other types of vehicles that will make a fleet greener and increase its efficiency, adjusted to the owner’s needs.
The optimal vehicle itinerary based on the charger type and proximity.
Suggested infrastructure upgrades
The number and type of charging stations that will work best for a specific scenario/delivery.
Who and how to train to get the most out of the available electric cars.
How much a business will be able to save by making the transition.
The difference in emissions between the current fleet composition and after going electric.
Overall EV-viability score
A general score that reflects how prepared a fleet is for the transition.
In recent months, fuel price spikes combined with high vehicle prices, depreciation, and inflation have led vehicle total cost of ownership to rise. In the USA, the average cost of operating and maintaining just one car topped $10K in 2022, growing by 11% compared to the previous year. In other countries, like the UK, this rise is not as pronounced, but still, individual expenses related to owning a car, like parts’ or repairs’ prices, have increased by up to 90%!
In these grim circumstances, the mobility sector seeks new ways to mitigate the rising costs. As you expect, the answer once again lies in data.
TCO is a complex mix of many values and operational expenses. Let’s dissect it to see how mobility data science can address each component.
Mobility analytics can assist fleet owners in making data-informed purchasing decisions by helping to assess car lifespan and depreciation. For that, the technology combines a variety of factors, including consumer trends, records of past car performance, brand reputation, component wear (for used cars), etc.
Direct operating costs
Automotive solutions can bring operational savings across the board: predictive maintenance will help reduce servicing expenses and optimize part usage; route planning will help save fuel (or energy) and eliminate unnecessary tolls and parking fees; car usage analytics will identify any vehicle misuse. Data analytics can also inform larger initiatives oriented at cutting costs, such as transitioning to EVs, as we explained above.
Indirect operating costs
Data on vehicle usage and occupancy can be analyzed to optimize the deployment of vehicles, ensuring that they are being used as effectively as possible and reducing the total number of vehicles needed. Data analytics have a range of uses in fleet management, from smart performance monitoring to route optimization, which can also help drive operational costs down.
We have an entire section devoted to residual value optimization, but in short, fleet management analytics use sensor data to evaluate how the usage of the vehicle affects its condition. Then, you can use that information to optimize maintenance and instruct your drivers on how to care for vehicles.
The tricky part about optimization is recognizing the areas that can be improved. Luckily, finding out what you don’t know is exactly what data science does best. By plugging fleet management analytics into the telematics data, fleet managers gain access to a new source of information to work with to increase the efficiency of operations.
Data science can be particularly useful in the key areas of fleet and delivery management.
How To Optimize Fleet Efficiency With Data
Increase time on the road by preventing accidents and unexpected failures
Enhance Route Planning
Pick the optimal itinerary thanks to data-driven insights
Avoid wasting fuel and time to improve productivity
It’s estimated that vehicle downtime costs fleets between $488 to $760 per car per day. These losses could be averted by adopting a data-driven approach to preventing unexpected breakdowns and accidents—the two main causes behind car downtime.
One way to achieve this is by departing from the traditional, condition-based vehicle maintenance models that rely on regular servicing. Instead, mobility businesses are adopting a preventive model that uses predictive analytics and ML algorithms to select the best time to schedule maintenance.
This is done by constantly tracking filters, battery life, tire pressure, or oil levels and having automotive analytics software use that data to choose the right time for repair, refill, or replacement. All this data is then collated with factors like driver behavior, location, and road infrastructure to identify further car health improvements such as driver coaching or rerouting.
Regarding accident prevention, mobility analytics allows fleet managers to understand driver behavior patterns and get a bird’s eye view of driving safety. Based on driving style trends pulled from cars’ telematic devices enhanced by external factors (weather, road surface, traffic, etc.), DSaaS software informs managers who drives responsibly and who needs driving coaching. It can also provide them with specific safety insights to minimize the risk and maximize time on the road.
Route optimization is the holy grail of logistics. That’s because selecting the right itinerary can save time and fuel, reduce car wear, and improve customer satisfaction. It also has the added psychological benefit of reducing driver stress—with more time to perform a task, they’re more likely to perform it well and safely.
But fleet managers know it’s not just as easy as picking the shortest route. They must consider traffic, time of day, fuel usage, weather, infrastructure, and carload, among other factors.
An automotive data analytics solution can process all that information in real-time, allowing drivers to sync current locations with delivery tasks, match carload to cargo, optimize tanking or charging times, and more. In a broader context, data science helps review common delivery routes and come up with improvements.
Although it’s impossible to keep your fleet fully productive at all times, excessive idling (i.e., stopping the vehicle with its engine running for five minutes or more) should be eliminated. It means wasting money on fuel and time a vehicle could spend more productively, essentially burning your fleet’s resources.
Automotive data analytics addresses idling by looking at how long the vehicle goes idle, where, and how often. Based on that knowledge, the algorithm detects idling patterns and suggests the next steps. This could be setting guidelines for drivers, implementing a reminder system, or eliminating behaviors like “warming up” the engine and leaving it running during unloading.
In the case of shipping companies, the key to increasing efficiency may lie in focusing on the final stage of the parcel’s journey. After all, the cost of last-mile delivery constitutes 53% of the total costs of shipping on average.
AI-based mobility platforms can plan the most efficient routes for every shipment based on average travel times, location of the distribution center, current traffic, car capacity, and other factors. With that info, the algorithm can also accurately calculate delivery times, enhancing planning. Consequently, shipping companies can dynamically adjust pricing to specific areas and circumstances.
In 2021, scientists from Kingston University and the University of Western Ontario published a study that discusses a predictive model capable of extrapolating driver decisions based on distance from the center of the traffic light and its position, the driver’s gaze, and head movement. The model can predict the maneuver 3-4 seconds before the car reaches an intersection.
Based on behavioral data, solutions like that show how much driver behavior can tell us. Luckily, if you want to tap into that knowledge, you don’t need to wait until these models are widely available. Insurers, commercial fleets, OEMs, and other mobility businesses already track behavioral data and use it to interact with drivers.
Many organizations collect useful data that allows them to learn more about their drivers—examples include location, dashcam recordings, or IoT records from connected devices. With the use of fleet management analytics, all this information can be transformed into insights and tools for engaging with drivers.
The primary use of these tools is coaching. Interacting directly with drivers through apps and dashboards gives many new possibilities to promote good driving practices. For example, one can implement an incentive program with rewards for the best-performing drivers. Gamification elements like rankings may motivate more competitive drivers to improve their skills and build a culture of healthy competition.
Another important feature is visibility. “Safe driving” is a vague term and can be understood differently. AI models condense a range of data points into a single driver score, presenting you with a great metric for evaluating and comparing drivers’ performance. Combined with other behavioral metrics, dashcam recordings, and statistical reports, you’ll get the full picture.
Then there’s the communicational aspect. Telematic-based analytics allows for tracking driver activity in real-time and sending automatic warnings when hazardous behavior (e.g., speeding or using a phone while driving) is detected. On the other hand, it’s just as important to praise drivers for making progress. For example, reflect on any improvements in the daily brief so that your drivers know they’re doing things right.
Overall, driver apps and similar solutions enable fleet managers to promote a responsible driving style and maximize safety proactively.
So far, we’ve talked extensively about how mobility organizations can leverage data analytics and science platforms to be more effective, profitable, and successful. Last but not least, let’s analyze the priceless contribution data science makes to enhance road safety and protect the health and well-being of all traffic participants.
Accidents can have many causes: bad weather, poor road conditions, misguiding signage, reckless driving style, heavy traffic, terrain, driver’s concentration, topology… All those factors may seem random, but in the end, most of them can be expressed through data—data that can be analyzed in search of patterns and potentially life-saving conclusions.
The need for new ways of improving road safety is still very much there. For instance, though many Western European countries have made tremendous progress in reducing the rate of road accident deaths over the past 20 years, mortality statistics remain high in many Central European states. The situation is similarly dire in the US, where traffic deaths hit a 20-year high in 2022. Globally, it’s estimated that 1.3 million people die in traffic accidents each year.
By analyzing common risk factors causing road accidents, mobility analytics can provide organizations from the mobility industry with insights that will help them reduce these numbers. To do that, data-based platforms process various factors affecting the situation on the road.
First, there are contextual factors, i.e., external circumstances out of the driver’s control. These include:
Road quality and topology
Surface quality is important, but how the road is situated can be an even bigger problem than potholes. Intersections are infamously known as accident hotspots; over 50% of fatal and injury crashes in the US occur at or near them.
Road safety is greatly influenced by the volume of traffic, who uses the road (drivers, cyclists, pedestrians), and their behavior. For instance, congestion increases the chance of bumps and non-lethal accidents.
Road segments that lack proper or no signage put everyone at risk and are particularly dangerous for pedestrians and cyclists. Worst of all, misplaced signage (e.g., temporary construction site signs) can obstruct the driver’s field of view.
Rain, fog, snow, side winds, and other weather factors affect the driver’s ability to control the vehicle and interpret hazards. Even seemingly positive conditions, like heat or sunny weather, have been linked with accidents by diminishing drivers’ focus or giving them a sense of overconfidence.
Realistically, governmental transportation agencies are the only organizations concerned with mobility that may influence some of these factors. Although exerting that influence can be difficult and time-consuming even for them, optimizing road infrastructure and surroundings with these risks in mind would improve safety by a long mile.
Apart from the above contextual factors, mobility analytics also factors in behavioral aspects to assess the risks and suggest improvements. These are, for example:
Manual, visual, and cognitive distractions are common causes of accidents, many of which are fatal. It’s estimated that in the US, nine people die every day in traffic accidents caused by distracted drivers.
It includes speeding, tailgating, ignoring signage and light signals, or erratic lane changing.
Now, how can fleet analytics software turn all this information into actionable road safety insights?
Predictive modeling techniques can be used to analyze historical data on accidents, such as the incidents’ time, location, and circumstances, to identify high-risk areas and predict when and where accidents are likely to occur. Public services can deploy resources and target interventions more effectively with this information.
Data from traffic sensors, cameras, and other sources can help identify traffic flow and congestion patterns and areas of the road network prone to traffic-related accidents. Transport agencies can improve road design and traffic management strategies based on that data.
The analysis of circumstances and causes of accidents helps identify common factors and trends like the types of vehicles involved, the age and experience of drivers, and the road conditions at the time of the crash. These statistics can inform the development of targeted road safety policies and interventions.
Data collected from vehicles, such as speed, acceleration, and GPS location, can be used to analyze driver behavior and identify risky driving practices. This data allows fleet managers to create targeted coaching programs to encourage safer driving behaviors.
While municipalities, commercial fleets, and delivery companies use automotive analytics to prevent dangers and promote safe driving, it also presents unique opportunities for the insurance industry. Based on the behavioral and contextual telematics data analysis, insurers can develop new insurance models to reward responsible driving.
Traditionally, insurance providers priced policies based on static factors such as car make, year of production, model, and condition, as well as the driver’s age, place of residence, and claims history. While these metrics affect how likely a driver is to engage in a road accident, this model completely ignores all the risk factors listed above.
Telematics data and mobility analytics enable insurers to take a new approach to policy pricing: usage-based insurance (UBI). The three main UBI models account for various aspects of car usage.
In this model, the price or premiums is based on the mileage. The basic idea is that drivers who frequently cover long distances are at a bigger risk of accidents. Another factor is fatigue and loss of concentration due to insufficient rest.
Here, pricing is calculated depending on driving style patterns. The algorithm considers how often and in what circumstances a driver performs risky maneuvers such as speeding, braking, or swerving.
In PWYD, the price is determined by spatial data. The model compares vehicle location with contextual and behavioral data to assess the risk based on road type, weather, other route details, and the driver’s reaction to these conditions.
These UBI pricing models have a range of benefits for providers and policyholders.
For insurers, additional factors to consider mean they can evaluate drivers’ risk scores more accurately. They can also be more flexible with pricing and customize the offer to individual customers.
For policyholders, UBI is more than just competitive pricing. With the behavior-based model, drivers now have more influence over the value of their insurance. Though claims history and age are still a factor, usage-based insurance makes it easier for young drivers to earn lower policy rates.
But perhaps the most important is how these dynamic, data-driven policies can encourage safe driving. Of course, every driver wants to avoid accidents, but for many, a tangible incentive in savings can make a difference and convince them to drive more securely. By adopting UBI models and taking clues from data science, insurers contribute to improving road safety, one driver at a time.
When Herman Hollerith worked on his tabulating machine, he couldn’t have imagined it would eventually evolve into the computers we know today. Even though so much has changed since his days, we still use technology to leverage the most critical resource: data.
Today, innovation allows us to collect and process more information than ever, and we still keep finding new uses for data across all areas of human activity. The automotive industry is no exception: mobility organizations use data science and analytics to inform business decisions, optimize everyday operations, improve safety, and more. Most importantly, information is a crucial enabler of innovation—whatever new inventions the future holds; data will help us make the best use of them.
Motion-S has everything from automotive industry focus to experienced data specialists to help you turn raw information into valuable, profit-building insights. So if you are ready to embark on the data-driven journey, and find your inspiration to innovate and succeed, just as Hollerith did on the train, reach out to our experts. And let’s turn your data into profit.
Data science and data analytics differ but can be used together for great benefit by businesses from any industry.
This includes all kinds of organizations across different industries.
When it comes to mobility, various involved parties take advantage of data, including insurers, commercial fleets, transportation companies, shipping firms, public services, car rental companies, and more.
Hiring costs, talent pool, quality, recruitment process, as well as the dynamic nature of data science projects should all be accounted for when companies consider setting up an in-house data team.
Did you like what you read? Share it!
Are you ready to dive into mobility analytics?
Contact us to discuss your use cases and needs, or sign up for a FREE demo. See how impactful insights can improve your chances of success!
Would you like to see our solutions in action? Check our calendar and find the perfect slot suiting your schedule!
Test drive our API Suite for 30 days!
Tell us a bit about yourself, and we’ll get in touch with you in no time.