Is Data Always Useful?
“Data is the new oil” is a phrase that has been inescapable in the tech industry for the past decade. On the surface, the maxim, coined by supermarket loyalty card innovator Clive Humby, suggests that data is a valuable commodity in the new economy. At the same time, it also suggests that, just like crude oil, data is not particularly useful in its unrefined state. In the same way that oil must be processed into gas, plastic, and more to be practically useful, data too must be “processed” before it can deliver actionable insights.
In recognition of this fact, a massive industry has sprung up around taking unrefined streams of raw data and presenting them back to customers in intuitive and actionable ways.

Instead of showing a stream of numbers over time, Google Analytics and its ilk have moved towards enabling users to ask questions about the data in natural language. Questions like “How many users did I have last week?” and “What’s my month-over-month growth rate for the past 10 weeks?” are now presented front and center in the popular website analytics tool. The idea underpinning this shift towards a questions-centric user experience is that people want their questions answered in the easiest way possible—they don’t want to look at streams of data and have to hunt for conclusions.
Why Not Why?
However, while data analysis tools are increasingly excellent at telling you what has happened in terms that resonate with you, they are not yet advanced enough to explain why (and this is a really hard problem that lots of smart people are working on).
Consider the case of a website’s bounce rate, that is, the percentage of visitors to a website who navigate away from the site after viewing only one page. A data analytics tool will tell you that your bounce rate is X%, but it won’t tell you what really want to know—why people are leaving? This means that you need to do some experiments (effectively educated guess work) in order to see if you can move the bounce rate lower.

The general expectation in industry is that analytics tools lead you into an iterative process, always trying to nudge the metrics you care about up or down, never fully certain about why the people or things that you’re measuring are behaving as they do. In the case of bounce rate it could be the type of visitors you’re getting, seasonal effects, website copy, website colors, imagery, load speed, or much, much more. Again, you can always know what is happening, but you can rarely know why.
Thankfully, the cost of experimentation within websites and apps is low, and you can quickly get a better understanding of how best to move the numbers. Conversely, the cost of finding out why something is happening, for example, by surveying visitors to your website asking why they are leaving, is very expensive because people don’t like it when you do that!
In complete contrast to the web-centric scenario above, when it comes to reviewing environmental sensor data, knowing why something is happening is absolutely crucial. The cost of experimentation in the physical world is prohibitively high, and the cost of misinterpreting data about soil, air, or water quality can have very negative effects, including on human life.
To take a simple example, imagine you installed some air quality sensors and learned that the air around your building was bad. You might go to the trouble and expense of planting trees around the building to improve the air quality, only to later discover that the underlying reason for the poor air quality was environmental pollutants emitted from a suboptimal heating system that could have been fixed far more quickly and cheaply.
Really understanding why a given environmental state is being observed is essential before you set out on your journey to fix it.
People Power
So how do we move beyond just finding out what is happening in the environment, and move on to understanding why exactly it is happening? The answer lies in people.
The environment is so complex that sensors alone can’t tell you everything you need to know to make decisions about how to improve your situation. From talking to our customers, we’ve learned that the best way to really understand what’s happening out there is to empower people to contribute qualitative information on top of the quantitative sensor data captured by the Temboo platform. This dynamic leverages the fact that people are naturally excellent at answering questions that machines are not best suited to like making observations about the physical world. Even more importantly, people are very motivated to contribute to making the world a better place (as opposed to answering questions about why they are abandoning online shopping carts).

In the gif above, the graph shows a spike in air quality on March 15, from good to bad. This is interesting in and of itself since you can use this information in the moment to make decisions about outdoor activities, and over the longer term to form a macro view of air quality patterns in the area.
However, it doesn’t tell us why the air quality suddenly deteriorated. This means that you can be reactive, but not proactive about fixing the underlying causes. Automated systems can’t tell you why something like this happened. In order to do so they would need to understand pretty much everything about the world, and AI simply isn’t that advanced, and is unlikely to be for a long time.
Thankfully, figuring out and adding the context needed to understand why the air quality spiked is a pretty easy job for a person with local knowledge. It’s the type of task that humans find easy and are excellent at.
Back to the example above. When the air quality spike was observed, the resident immediately guessed that the cause could have stemmed from an incident on the nearby highway. A quick check of the local news revealed that a car fire had happened at exactly the same time as the air quality spike. The resident went ahead and added a note to the data stream for all to see, and transformed a “this is what happened” event into a much more value “this is why that happened” observation.

In a second example, a poor air quality spike early in the morning of July 5th is easily explained by a community member who observed fireworks being set off in close proximity to the sensor.
By building an environmental monitoring system that puts as much emphasis on empowering qualitative contributions by an engaged community as it does on sensors, and by leveraging people’s innate interest in helping improve their surrounding environment, tools like Temboo can do better than just telling you what happened, and help you really understand why.
Not only that, but monitoring the environment in this way leads to more engaged communities, a more stable monitoring system through volunteer maintenance, and upskilling opportunities for community members who get to work hands-on with emerging sensor technologies and social analytics software. This is better for you as a Temboo customer, and better for the people who live in the environment you care about.
Could you benefit from community-powered environmental insights?
Are you interested in learning more about how you can use Temboo to really understand what’s happening in the places you care about, and how to unlock the engagement potential in your community? Get in touch with us today to chat about how we can help you get started!
You must be logged in to post a comment.