Reimagining Analytics. Wait, what?!
Before we even start on rumbling about what the heck reimagining analytics is, I’d like to let you know that this article is about big data analytics hype and what (according to us) it actually is. First of, lets start with big data itself. Big data is in essence a new approach on storing vast amounts of data that has the 3 V characteristics (Volume, Velocity, Variety) and can be used for subsequent analysis. The concept is introduced around 10 years ago by the silicon valley big boys (Google, Yahoo) and is currently used by every major internet corporations around the globe. The promise or the hype of big data is to enable corporates/government agencies to get a 360 degree view on all of their data, be it structured (databases, spreadsheets, etc) and unstructured (freetext, socmed conversations, etc). Imagine what you can do with all those data, when it is combined in a single place, regardless of formats and sources (in our case, its analytics). Sounds awesome right?
Not so fast cowboy. With its perks and possibilities, big data also has its challenges. The notable challenges would be communication issues, presence of information silos & inability of sharing data across business division. Big data will never work unless the stated issues are fixed. The whole organization has to work together and have the same vision about consolidating their data in one place for everyone to use.
Ok, we got the drill and got the main challenges, so what? What if my company is still profitable without using this big data thing? What if I’m still doing alright by using my intuition for my business performance?
Connecting structured and unstructured data
Those are some good arguments, but hear me out now. Here’s the thing, what is the main purpose of every business? To obtain profit of course, but to obtain profit, businesses needs to do many things, and business performance analysis is one of them (and a pretty important one, too). The real question is how many variables are you guys using for your business performance analysis? What are you guys doing with those variables? And what about other data, such as social media, web, and the massive unstructured data floating around, are these counted as variables? How do we deal with these unstructured data?
That’s where the analytics thing becomes important. Now, if we can connect the data from social media and the organizations sales and operational data, we might find correlation between structured and unstructured data, i.e. customers comments and sales result, and basically find how to ‘fine-tune’ the business.
Awesome right? Yes indeed. So, next time somebody is explaining to you about data analytics, stop calling them geeks or nerds, smart is (and has always been) sexy, and they are going to be the hottest thing in town.
Imagine and reimagine things that you can do with your data
I’d like to share with you guys some samples on how fun analytics is. The first one is from the movie Iron Man 3 (ref:a), it’s a scene where Tony Stark is trying to figure out what is causing the explosion that left his friend & driver Happy severely injured in the hospital.
Tony stark asked his personal AI assistant (Jarvis) to pointed out explosion cases in the last few months after the occurrence of the villain (Mandarin) which has similar heat characteristics with the one Happy was in (>3000° Celsius). The result shows that one such event occurs in another place, which subsequently brings Tony Stark to visit the location in the following scenes after.
These movie scenes might makes one think that “oh, it’s just a movie”. Well, it’s possible, and is actually in practice as we speak. Those correlated and interconnected data is can be obtained with the aid of Graph Database. To put it with better understanding, I’ll try to show you what Jarvis provide to Tony in a simple network analysis which is used in graph database.
Here’s Tony with the assistance of Jarvis trying to have a better view of Mandarin’s attack.
and here’s how a simple network analysis explain it
Another example is this quirky article by Emil Eifrem (ref:b) (Neo4J founder) entitled “what can banks learn from online dating” which emphasize on fraud analysis. Emil stated that
“One of the features of first-party fraud is the exponential relationship between the number of individuals involved and the overall currency value being stolen. For example, 10 fraudsters can create 100 false identities sharing 10 elements between them (name, date of birth, phone number, address etc.). It is easy for a small group of fraudsters to use these elements to invent identities which to banks look utterly genuine.”
This presents both sides of the coin. In one hand, by involving more people, first party fraud is a promising approach for an organized crime. On the other hand, participation of networks of individuals is actually makes the job of investigation easier, however.
By using graph database, financial institutions will be able to identify these fraud rings through connected “social network” analysis (similar with what is described in the Tony Stark vs Mandarin earlier). This involves exploring and identifying any connections between customers before looking at their spending patterns (FYI — you’ll go mental if you tried to do this with a traditional relational database).
So, what’s next?
Big data is merely a new approach of collecting the ingredients for analysis, which is data. At the end of the day, what the users/businesses needed is analytical result of their enterprise data, to improve their business performance and/or enhance their service delivery.
Using analytics to find correlations between entities/values might lead us into new insights on how our data is connected to one another and how our business is doing.
Hence, don’t get stuck with just one approach with your analysis. Keep searching and dream about things you can do with your data, and reimagine your own analytics.
a: Iron Man 3
*Originally written in my LinkedIn Article Post.