Analytics Software

Fuzzy Logix CEO stays in databases to chase insights

Fuzzy Logix CEO Partha Sen is a numbers man. When not leading his data analytics startup, the India-born Charlotte, North Carolina resident attempts to call the results of cricket matches and other events based on the stats to hand.

Sen’s magic ball at Fuzzy Logix is analysing data from within the database – no extraction routines or additional storage. He holds that that direct-to-the-source approach is part of a new wave of analytics tools built for modern challenges where oceans of data from a bewildering variety of sources must be examined in close to real time.

“Conventional analytics is struggling to move this large volume of data,” he argues. “It gets memory-bound because nobody thought you were going to analyse so much data. Our USP is that you’re not going to have to move the data and all the models can be executed in parallel. Software is embedded in the data warehouse.”

“Over the last 20 years there has been a proliferation of data: through the internet, globalisation, digital financial systems... there’s more and more, a greater volume of transactions, debits and credits to be handled. Cash is becoming the new parallel economy now. People are building hundreds of thousands of models. Retailers are trying to make a forecast for the supply chain of every single product but conventional analysis only runs one model at a time.”


The new black

Predictive analytics is the “new black” in technology circles but Sen prefers broader foraging ground.

“An analyst comes into an enterprise and he works on data and he’s going to use a basket of techniques: predictive modelling is part of it but you also need the inferential, using historical data. There are many models and sometimes it’s not just predictive. Predictive is a buzzword and people latch on to it.”

He also queries all the attention piled on Hadoop.

“Databases are going to stay, in spite of the hype you have with Hadoop,” he says. “Hadoop is popular because of the cost factor but you have to give up on some performance. I wouldn’t be surprised if on ultra-high performance you have some other type of database like [the technology used by] StreamBase.”

So far Fuzzy Logix can point to big-name customers such as Tesco and Siemens. It has raised only a modest $5m in early funding but Sen says the company is self-funded and has angel backing but there will be more cash raised “like filling your car with petrol”.


A whole new world

Sen doesn’t sound like he’s in a hurry to exit and says that it could take a while for Fuzzy Logix to fulfil its true potential as developments like the Internet of Things accentuate the need for new approaches to processing, querying and storing data. He confesses to having had a “healthy disregard for the classroom” while at university but he was always fascinated by numbers and analytical models and he believes we are entering a whole new world.

“I think the opportunity is understated,” he says. “You’ll see generations of data. Data generates more data and I don’t think anybody has taken that into account. If you look at China it’s basically a manufacturing hub but I think services will pick up at some point there.”

Each big change in what we can do with data will require new approaches, Sen believes, and he cheerfully admits that even his company’s current technology won’t be adequate 10 years hence.

That’s very likely true and for all the obsessing over modern analytics proving the answer to the world’s great questions, it’s still fiendishly hard to get the answers right on a consistent basis. Shortly before this article was drafted Sen used his blog to predict that the Carolina Panthers would beat the Denver Broncos by 27-10 but instead Denver took the Super Bowl. You can’t win them all.


« Typical 24: Karl Mendez, CWCS Managed Hosting


Perspective: A woman in tech in Palestine »
Martin Veitch

Martin Veitch is Contributing Editor for IDG Connect

  • twt
  • twt
  • Mail


Do you think your smartphone is making you a workaholic?