Security cameras capture footage used by computer vision software to analyze scenes and summarize insight report.
Since the inception of the “quantified self” and corresponding hardware devices, I have been curious about what people are doing with the data they collect about themselves. Fitbits, Apple Watches, and step tracking tools provide individuals with insight about their physical activity. With the increase in data generated about individuals, I tend to wonder how this information translates to groups.
Businesses could use data gathered in their establishments to better understand their customer behavior.
Hardware companies have written interesting studies analyzing the data their platforms and devices collect, but this information is restricted to the company’s chosen points of interest. Smaller establishments, such as restaurants or commercial stores, do not have the benefit of gaining insight from the increase in these data points.
I think there is an interesting way to think about the other sensors that are newly available for analysis, specifically, the sensors that have always been around, but have not been fully utilized.
Tracking trends in public spaces is now available to all
In years prior, sonar, lasers, and similar one-use sensor technologies have been required for computing the physical world with data. Tracking the traffic trends in public spaces or tracking the foot traffic in commercial establishments have previously required specialized hardware. The specialized hardware was available to cities and large commercial ventures, but again, not for smaller scale commercial ventures.
Security cameras in restaurants, cafes, and bars could be used to monitor customer behavior.
Software advancements now make common use digital cameras into complex sensors. Using software that processes images, the content of an image can be quickly analyzed for recognizable objects and human behavior. Computer-vision based software can be used to reconstruct scenes using networks of cameras. This allows for common security camera infrastructure to be used for real-life google analytics like analysis.
Data gained from analyzing customers could be used to gauge customer loyalty or predict behavior.
Human activity in a common space, monitored with a digital camera, can be used to interpret common movement into analyzable data points. Further, in commercial establishments, transaction data can extend these reference points to understand how a brick-and-mortar business operates. I assume this information would not provide revolutionary insights for a business owner, but would be a great reference over time for change in business activity. It’s reasonable to assume this data, over time, would help to foresee trends in business growth or decline.
In the immediate, this kind of monitoring and analysis is uncomfortable. I’m curious to see if our expectations and cultural norms will change with time. Already, our activity online is analyzed and used as a commercial product. The click-trails we leave behind are aggregated and used to better target ads. The search behavior, communication patterns, or social graphs we create offer greater insight into the kind of person we are than most of us would like to admit. This monitoring of behavior is something people have become both more sensitive to and less aware of.
How much right does a business have in tracking their customers? Does it make a difference for the customer if they don’t know they are being monitored?
Given the changing times, I wonder if brick-and-mortar businesses would be comfortable using their existing infrastructure to gain deeper insights in to their business. Further, I wonder how customers would feel knowing that their activity is being monitored and analyzed for the benefit of a business.
This technology is already being widely used in public places. Companies like Sidewalk Labs that offer free wifi using publicly placed kiosks outfitted with high quality cameras, which are undoubtedly being used to monitor public activity. I’m curious if this kind of monitoring and analysis would have required a different approach using older hardware solutions in the past.