Forget Everything You Think You Know About Big Data Access.


Recently I sat down with my family to watch a DVD – not something we do often with all of the streaming options available, but I digress. Our only DVD player is an old Sony PS3 – so I started to fire it up, change the TV input source, and grab a game controller to start mashing buttons until I figured out how to get it to “play.”

My 9-year-old looked at me like I was crazy: “Why don’t you just use the TV remote?”

Me, seeing this as a teachable moment, scoffed and said: “You can’t just use the TV remote, son – see you’ve got to control the PS3 with this controller, it’s not the best interface but . . .”

While I was talking, my 9-year-old picked up the TV remote and hit play, and MAGICALLY the PS3 started playing the DVD.

“What? How did you do that?”

“It just works, Dad.”

So evidently my 9-year-old, without all my preconceived notions about how frustrating audio/video tech can be, lacking the years and all the ways I’ve learned to work around the limitations, had picked up the remote and figured that pushing play would just work. This is something I didn’t even think to try because I approached the situation with my years of adapting and accepting the limitations of audio/video technology.

We are at the same point with data analytics, and this is why I’m excited to be a part of Molecula. We are conditioned to look at large data sets and assume that we can get something out of them, but only at the cost of waiting hours, even days to successfully run a query against it. 

It’s time to change that mindset!

I was in a meeting recently with a large technology company. They had this really cool interface that was built on a huge source of data, which they then sliced and diced in batches, overnight, every night, so that the next day they could provide a very specific and very cool visualization of that data. The technology was cool and the insights they were able to visualize were incredible!

The Molecula team started what-if-ing on other uses for this data set, but since we could only see the results of the pre-processing of the data, we started asking probing questions about the actual data source: what’s in it? Can you correlate to X? Does it include this or that timestamp? And so on. The answers came back – 

“Well, I’m not sure.” 

“We haven’t considered that.” 

“All we can really do is work with the output of the overnight processing.”

The data set was so large and unwieldy and the company was so focused on the single output (which again, was very impactful!), that considering other uses caused their heads to spin. It would take too much work to get those kinds of answers. The processing alone could take days to weeks. It’s understandable why they hadn’t considered all of the value they could get from this data given their acceptance of so many technological limitations. However, it’s a shame and a mistake that could be costing this company incredible data value and insights that could quite literally transform their business. 

What if you had the ability to speed up data processing by a factor of 10? 100? 1000? What if overnight batches became real-time, even UI-time output? How would that change your world of possibilities with your data? What ideas have you not considered because it would just be too hard to process with your current technology?

Molecula’s FeatureBase will obliterate the limitations you’ve probably long-accepted about your data accessibility and infrastructure. 

Like my 9-year-old son picking up the TV remote and opening my eyes to what’s possible, Molecula is saving our customers from the current pains of the status quo of data access with its feature extraction and storage platform.

Want to hear how we do it? Let’s schedule a time to start the conversation and help you maximize your data ROI.

How can we help you re-think your big data mindset? 

About Molecula’s FeatureBase:

FeatureBase enables real-time analytics and AI initiatives by automatically extracting features from data and storing them in a concise, optimized format specifically designed around making features individually analyzable and accessible. This feature-first approach makes model-ready data reusable across the organization, without the need to pre-process data.

To learn more about Molecula:

Check out our products and solutions.