Conversations with Data: Unearthing the Value
Now, Let’s look at how applying context and looking at data in relation plays a vital role, in unlocking the magic!
In our last article available here, we looked at how a conversational mindset can help us understand our data better.
Unearthing the value of Data
When we were working on a Kaggle project to determine if a driver was drowsy or not, the pressure exerted on the accelerator by the driver was just one of the input datasets. By itself, it may not mean much. It actually depends on the driving practice — someone like to rev the engine every now and then, someone has a steady acceleration and so on. But if I juxtapose this with any inputs on lane infringement, I may have a scenario where the driver may be drowsy. I add in a third element — the steering wheel movement. I can now determine, with reasonable certainty, that the driver may have fallen asleep if these values are beyond the thresholds used in the model!
What was an independent data point achieves exponential value when seen in relation to other data points. Finding out these possible relations and exploiting them is to release the magical power of data!
Data’s true value, hence, can be compared to an iceberg. With the tip that is visible, you can still derive conclusions and make decisions. But, if you wanted to really make smarter and better decisions, you would have to dive below the surface and check out the bulk of the iceberg.
The true value lies in laying out the vision for data and looking at it from a futuristic perspective, rather than binding it to immediate need at hand.
Here’s another example:
A Hemoglobin count may not mean much by itself.
Now when I couple it with other key counts such as that of RBCs, Platelets… and map it over time, I get an amazing plot of how someone’s health has been going up and down, with relation to the overall treatment! So, in one shot, I know what worked and what worsened the condition.
We all have heard of organizations rushing headlong without a bigger picture understanding of data. They find out too late that embarking on a transformative journey would require extensive work on the current data and structure to enable future-readiness!
An ideal example would be companies crawling for data across millions of websites and getting a lot of data in to power a product, only to realize that a minor enhancement to the product would need the crawling to be performed again to get the additional piece of information! Guess what, many a time the data isn’t available anymore leading to the collected data going waste as well!
- We have organizations such as Amazon, who while driving the technology behind AOL’s e-commerce site, understood what were AOL’s customers looking at and what they were buying. This information was promptly used to build and better their recommendation engine! AOL, despite running the commerce channel, did not realize this aspect of data usage!
- Google not only used the street view data for its ubiquitous Google maps, but also in the development and training of its self-driving cars!
- Netflix today shows me a lot of top 10 lists — this is an alternative use of the data that it once used to drive my recommendations! To give me a sorted list and inculcating a fear of missing out on a trend, Netflix makes me watch more!
The incremental cost in such cases is pretty low, but with a little thought, the extensions or new product lines can be runaway successes! All it requires is a thought through data strategy to ensure value!
Summary:
Like colours coming together to form an impeccable rainbow and the myriad of possible colours, Data gets its power with context, with relations.An innocuous data point may achieve significant importance when placed in the right context! Understand the context, look around! Don’t stop at the surface!
Next Article in this series: Letting Data drive your transformations
