One of the best ways to measure mobility data is through humans themselves. Turns out — we’re great sensors for this kind of thing. Once you’ve anonymized and segmented mobility data taken from our movements, the wealth of information is staggering. Dr. Arturo Amador of Capgemini works with this exact kind of data, gleaning insights into the way human mobility drives the economy and provides a breadth of data not found in traditional IoT sensors. In his talk for ODSC’s 2019 Accelerate AI, “Big Data and Mobility Analytics: What Can We Learn from the Way Things (and Humans) Move?” he gives us a glimpse into the power of big data and mobility.
Cities as Economy-Drivers
Cities are living organisms. Analyzing human movement throughout a city produces a graphic that (very poetically) mimics a heartbeat. This movement is tied to the economic drivers of the city — things like work, transportation, and residency.
Aside from massive disruptions – celebrities or political figures arriving in town, for example – these movements are reasonably regular. Cities provide value and survive through these self-organizing means.
Another source of movement is the transportation of goods. All of a country’s top industries rely in some part on transport, of which, cities are major drivers. Mobility is a fundamental part of life, and nowhere is this more apparent is in a city.
What this movement tells us is that value is exponential. As cities form and organize, they provide impacts that often survive from civilization to civilization. It’s no accident that cities like Rome or Mexico City have survived multiple iterations of civilization, and this mobility data could finally shed light on exactly why.
On a business scale, cities are using mobility data to create revenue for the city itself. In Boston, for example, driving a personal car into the city for work every day could cost as much as $20 in tolls while electric vehicles and public transportation pay no tolls. These policies create not only revenue but deep behavioral change over the long term.
Location Data: From Static to Dynamic
One of the most significant changes in mobility data is IoT itself. In countries like Norway, plentiful, rich data from many IoT sources allow us to see into the heart of the city. Norway, and the rest of the world, is monetizing this data through tourism, traffic monitoring, public transportation, and public health.
This surveillance is anonymous but provides vital insights. For tourism, this data could show patterns in travel habits based on nationality without paying for expensive personal monitoring and offers more data than simple static methods.
Artificial Intelligence is also driving this new era of dynamic analytics. It’s able to find patterns and interpret big data in ways that previous processing wasn’t able to. Now, AI is allowing us to look deep into the mobility of persons throughout cities and the world for real-time insight.
For business, this technology also provides insights into the movements of consumers. The data is available anonymously, but it provides critical data for how people move through a store, for example.
The Privacy Challenge
Privacy in this data collection is challenging, but companies are finding ways to comply – privacy by design. Each has benefits and downsides, but all can protect the privacy of individuals without explicit consent.
- Encryption: data at rest and data in transit.
- Masking: Masking techniques such as hashing could be suitable alternatives to encryption where possible.
- Extrapolation: Preserves privacy but could introduce some uncertainty
- Path obfuscation: Pseudo-random noise can preserve privacy by preventing re-inference (but decreases accuracy)
- Aggregation algorithms: Strengthen privacy frameworks by avoiding exposure of individuals — think k-anonymity or t-closeness
Without the combination of these privacy measures, its still possible to recover data and reidentify individuals, putting companies in direct violation of compliance measures like GDPR.
Merging Data Sources
When you begin merging these data sources, IoT plus local AirBNB data, for example, it creates an even more dynamic picture of the revenue and impact mobility brings into a city. For example, the data from cell phone movement combined with sentiment analysis of AirBNB reviews provides well-rounded insights about tourism.
Companies are also using wastewater analysis to plan anti-drug campaigns. AI can analyze anonymous data from wastewater based on mobility patterns and drug use. This data can provide realistic data about drug usage that could go against common assumptions. For example, AI revealed that drug consumption goes up per person in the summer. On the surface, it seems that drug consumption goes down, but in reality, with a lower population and higher numbers in wastewater, we find different results than what we assume.
Big Data Tech and Maritime Navigation
Finally, this data is also possible to extrapolate for things. In maritime data, for example, you can predict port demand and saturation. You can also use this data to predict travel times based on factors such as weather, routes, and time of year.
With this data, route optimization is possible. Because most countries rely heavily on transportation for economic drive, this capability is a huge part win for AI. It makes the production pipeline more efficient and increases value for port cities and countries exporting goods.
Making Better Use of Mobility
Dr. Amador believes that this data is a massive boon to the impact and value of cities, allowing countries and businesses to monetize through logical and practical data interpretation. As the world moves towards real-time insights, Dr. Amador believes that we’ll be able to comply both with privacy regulations and build better, more robust data sets through other means.
Take advantage of ODSC’s learning resources by watching past talks on some of the hottest topics in data science and AI.