Big data is growing. Shocked? Sure, you must have heard about it again and again, to the point it sounds like another lame statement. What you will be amazed to see is how few are actually doing anything about it. The fun begins when we start digging deeper into what constitutes to be the real problem, is it the capability to handle data, data channel, query data, data prep or the big daddy, analytics. Hmm.. How about calling it Big Analytics.
Let me put it in another way: “Big Data is a Big Analytics Problem”. Sounds strange? – per CIO Insight “81% of IT professionals around the globe agree that better analytics is central to the enterprise’s big data problems”
Before I get to the point in the talk, let us take a quick detour to current bottleneck that analytics world is facing, that could potentially restrict it from growing:
1. Human problem with Analytics: I know you must be already feeling the connection here. It’s you, it’s me, and it’s us. We all love our analytics to death and we hate sharing it. We share it with people we have established analytics intimacy with. Your work buddies, team, boss etc. But there is a problem. What if the main gravy that you could learn from stays outside your reach? How will you ever find out what others are doing? I am sure you must be thinking, I am the cutting edge guy. Sure, I feel that all the time as well, but what if, there is something out there that we could learn from. What if you are pushed to stand in front of datasets that some other guy in some other team has already mastered, how will you scale. Problem, right? Yes, completely human driven analytics starts to fall apart from this point.
2. AI problem with Analytics: Now let’s get to artificial intelligence, alias machine-learning, alias that si-fi s#$t. Now don’t jump on me yet. I agree that AIis never meant to completely out take analytics. We have created a restricted container for our machine learning AI friends. Their task goes in, results come out. They are patted on their backs, life moves on. But there is a problem, human deal with most of pressing analytics that is directly impacting the businesses. So, AI alone will not sustain, and there will always be a need for human analysts, to make sense of AI friends. So, sci-fi is incomplete to handle growing demand as well.
3. Politics, perception and whole 9 yards: Yes, I love this part; I have gottenyelled at and a few times shown the door. Many believe analytics is a sacred entity and guarded with utmost care and should obey to company politics and perceptions. Smelling funny? Don’t worry, if things still appear rosy. Our brains are wired to accommodate the politics, regulations and compliance. I am not against it either, but having your analytics capabilities enslaved by some paralytic limitation could really restrict you from growing. So, make sure you understand what the restrictions are and how much you could let it float. Remember, there is always learning across that closed wall.
4. Current Deliveries Vs Long Term Vision: Yay, appraisal talks! We are all compensated for our current and short-term deliveries, that plague analytics as well. Who will be evaluated in pursuing something that may show its color in couple of year if not the next? Yuck! Believe it or not, it is what is another bottleneck that current analytics standard face. We need to build analytics strategies and models accommodating changing business needs and environment variables. Scalable methodical analytical models will help you tame the growing data needs and help in grasping bigger picture which most of the time is missing from the equation. All you need to do is combine long term and short term analytics and give your analytics strategy a hard and long view. Maybe, you could find a magic sauce that will help your business stay afloat on awesome analytics.
5. We are doing it, look at our BigData tools: First, Bravo! Getting infrastructure right and committed to your bigdata need is a huge step most of the laggards are still crying about. So, I commend you for crossing that chasm. But, remember, bigdata tools are not substantial to solve your big-data needs. In fact, bigdata tools and capabilities will make available more data and cleaner data at shorter durations for you to play with. Which in no ways mimics the speed at which at we do analytics. Our analytics are pretty much enslaved to our current needs. There is a strong chance that things will fall through the crack and might never get noticed. Current tools and capabilities are not talking about urgency of revamping your analytics strategy.
Almost done with my rant. So, what’s the conclusion? Have a strategy that will help you grow your analytics to handle the growing data. Just by buying boxes, infrastructure layer and other toys to handle your data deluge might not provide a complete and rounded solution. You still have to deal with your traditional analytics and change it to reflect ROA / ROI on your awesome big-data toy investments. So, up your sleeves and get cranking on drafting a scalable analytics framework. Want some idea? Hush.. You heard it here first; let’s call it “collaborative analytics [™]”. So, what is collaborative analytics? If you’ve read the blog, you already know it. Are you still curious? Wait for our follow-up blog. Till then happy analyzing!
Latest posts by v1shal (see all)
- Surviving the Internet of Things - April 22, 2014
- Using Topological Data Analysis on your BigData - April 15, 2014
- 3 S for Building Big Data Analytics Tool of the Future - April 14, 2014