BackHow tech experts are tapping big data
Here's what experts across sectors had to say at the 2016 edition of Mint's Enterprise Technology Summit, held in Bengaluru on 20 May
Premium
By 2020, the world would have generated around 40 zettabytes of data, or 5,127 GB per individual, according to an estimate by researcher International Data Corp. By that year, researchers also expect about 50 billion devices and things to be connected, making the Internet of Things era a firm reality. Here’s what experts across sectors had to say at the 2016 edition of Mint’s Enterprise Technology Summit, held in Bengaluru on 20 May.
Deriving power from Big Data
The power of big data is in the multi-directional insight it can provide. The story about big data, is that it is moving from descriptive to diagnostic (why that happened) to predictive (what will happen) and now to prescriptive. The cost of doing something wrong is too heavy.
When we think about IoT (Internet of Things), it is connected things. In reality, IoT can even be connected cows that show us how to improve farm productivity. The opportunities are large. As researchers, we focus more on technologies, and then figure out what it can solve.
Fortunately or unfortunately you don’t have an option of not doing it (big data analytics). You either lead it, or follow, or you won’t be there. The same thing was said about the internet a while ago, you rarely see businesses not being on web today and in IoT’s case, the value is even more powerful. It is just not a device.
Confluence of IoT and big data: Big Data came about not because of IoT, although, nowadays we can’t think about IoT without Big Data. Big Data came about primarily because of transactional flaws, emails, audio, video, even that itself had traditional databases. The well understood paradigm of SQL-based data bases was no longer valid. We had to rapidly think about how do we manage billions of pieces of data. IoT became real because big data already existed and it started leveraging it.
Three Vs—volume, velocity and variety—drove big data technology to the current stage. Variety is ability to integrate data and understand and leverage them. Then there is veracity and value. Veracity is how do we ensure that the data you are getting is authentic. And in the IoT world this is a much larger challenge.
Big data is an enabler, and big data has to grow for IoT to be real and the big industry players are already on to it. —Shriram Revankar, VP and fellow, Adobe Research Big Data Experience Lab.
What IoT means for cars
We operate on three principles: accepting no limits, alternate thinking and driving positive change.
An example on alternative thinking is how we launched car diagnostics through a smartphone. Generally what happens is that the diagnostics are done through a laptop which you connect to the car and then go through some kind of protocols. But recently we have enabled the smartphone to diagnose by connecting through Bluetooth and we are the only OEM (original equipment manufacturer) to do so.
In the context of IoT in the automotive sector, there is a dramatic increase in the overall content of the sensors, micro controllers and the computing power available in cars.
For example, when we launched Scorpio, a sports utility vehicle, in 2002, we hardly had any Electronic Control Unit (ECU)--a system that controls one or more of the electrical systems in a vehicle, on board, and there were hardly any sensors. If you look at it today, we have about 12 ECUs and almost about 80-90 sensors, and more than 1,000 circuits in the car. At any given point of time, 184 MB of data is getting exchanged through a Controlled Area Network every hour.
Our flagship model, the XUV 500, has more than 20 ECUs and over 1,500 circuits. When computing in cars takes the next step, cars will get more and more connected to the external world. We will see them communicate with other cars, with traffic signals, parking lots, or any other external service providers.
And IoT can be a big differentiator for companies in the automotive sector and play a big role in every part of the product life cycle. —Srinivas Aravapalli, principal chief engineer and head of vehicle systems at Mahindra and Mahindra Ltd
Connected factories
Bosch is a things company. We make a lot of things, products, and we have a goal that in the next year, all our products will be connected. That means we will be introducing sensors and measuring what our products are going to do. We also have some very aggressive targets to achieve.
Our CEO has mentioned that if the factories’ processes are connected, you can achieve close to 30% improvement in productivity. That’s massive from where we are today, looking at the number of plants we have, over 200 plants in more than 150 countries. The overall savings and the amount of productivity that you can generate is staggering.
One of our plants in Europe has been declared as the connected factory. If you look at the numbers, close to 5,000 systems have been connected together which means that they are talking to one another, there’s a good traceability of goods, smooth flow of operations and the overall advantages that technology has to being in is now being realised. This is part of our connected factory project which is expected to bring in a sea change in the way the digitisation of plants work.
Another important thing to remember in IoT is that by solving a simple problem you will not get benefits. If you make a machine automated, if you make an instrument to spew out data you will not solve the problem because eventually you will hit a bottleneck.
So the whole purpose of IoT should be to look at the bottlenecks and introduce intelligent sensing in those bottlenecks and clear the bottlenecks.
How they reduced the scrap wastage in manufacturing
For example, by looking at close to 600 variables in 16 stations in the assembly line and looking at data over a period of time we have been able to bring down scrap by 67%—that’s massive in terms of material usage. If you put a cost to the material, you are saving close to 2.4 million euros annually.
So, this is where it’s very important to identify and look at your wastage. Where is it that you are losing money? Where is it that you can gain time and gain an advantage? IoT is to be implemented there to be able to get relevant advantages. —Uday Prabhu, general manager, Internet of Things and Product Management, Robert Bosch Engineering and Business Solutions Ltd.
Creating digital twins
To understand how have to position our businesses in the future we need to understand how we got here.
You start with one thing, like Amazon did with books, leverage data and you start learning how behavioural understand can be customised, learning from various individuals and driving business opportunities. From there, you move into adjacencies—now Amazon sells everything, including computing power.
The three paradigms behind this sort of transformation are platform, aggregation, and personalisation.
We, at GE, intend to do the same for driving efficiencies in the industrial sector.
If you look at how we maintain equipment, it’s primarily based on time based-maintenance. If you take cars, with the digital transformation, what we’re doing is collecting more info on each individual engine, information that tells you that in the last inspection, this particular thing was doing really well, you don’t need to do it in the next cycle.
This can transform your profit and loss—you can know from each individual asset how much money you can make—both for yourself and for your customer.
We build digital twins—technologies that model real world individual components, devices. For example, in a wind farm there are regulations that say you have to produce power within an error of 15%, but wind energy is not stable. So we built the digital twin of a wind turbine, and combining that with weather data we provided accurate predictions. We made two digital twins of two different turbines talk to each other and collectively they produce 5% more power, and offer 20% more profit. —Vinay Jammu, Technology Leader in GE Global Research, and part of the company’s Software Science Analytics Team.
Personalised retail with Big Data
Target is the second-largest retailer in the US with more than 1,800 stores. Target India was set up in 2005 and has about 2,600 members.
The 3Vs—volume, velocity and variety—started defining Big Data. When you were clicking on websites and traversing through it, we were generating a lot of data. It is present in industries including retail, banking, logistics, travel and transportation. We were always producing large amounts of data but it has become intense with multifold multiplying of devices. Data consumption has also changed.
On Target.com’s Cartwheel app, you get offers on your phone and can swipe the barcode at PoS (point of sale) and get discounts on that particular item. This needs a lot of big data analysis as to what items/offers you going to push, what is going to be redeemable, interest. Cartwheel captures your click data, browse behaviour, purchase history and look at other guests, see recommendations of the website and which you have seen is the people who viewed this also viewed this, people who viewed this ultimately bought this and this needs a lot of number crunching and Big Data trying to find commonalities and patterns to show you recommendations.
We have another functionality called Crush where we have turned the personalisation and recommendations on its head a little bit. You can “Heart" an item and from the moment you do this, there is a lot of on-the-fly algorithmic computation behind the scenes trying to find out if you like these kind of items and alerted when there is a sale on those items, you will get notifications. This is something you choose to tell us and we are able to analyse and predict other items you would like to redeem and know about.
Shopping will completely change with more connected devices because of IoT. The assortment and merchandise that is surrounding you also will become much more dynamic--no static prices and promotions and need based supply inventory.
With augmented and virtual reality, the experience will be different like the store already knows who are when you walk in. —Seemantini Godbole, VP, Digital Technology, Target India
Building smart hotels
Travellers and expectations have changed dramatically with technology in the last 15 years. Earlier they expected good coffee, good linen, AC and a pool but now in lobbies, everyone is connected and we need to adapt to those changes.
At Starwoods, where we have 1,300 hotels in 100 countries, we measure the Guest Experience Index. Until now, this used to be limited to high-paying clientele and it was based on historical data which a person at hotel a person used to feed into hotel. It wasn’t accurate and the personalisation was not very high.
So we launched an app about five years back, and we had an app for each brand of ours. We are able to provide our customer with the entire customer life cycle on this very app. You can explore, review, browse, book, it becomes the room key. This even supported the Google Glass.
In our design labs where we are building hotel rooms for tomorrow, there are close to 70 sensors capturing different sorts of data and continuously updating the room, customising the rooms. We are deploying intelligent algorithms and processing a lot of data to provide meaningful experience to guest.
You can customise AC, lighting, temperature, format channels, all before reaching the hotel. We even have a concierge robot called Botlr (robotic butler) that helps in allocating room and guides them into room, provides information around hotel. It even acts as a room service butler and serves up food to your room. The door lock senses the Botlr is there, it makes a call to the hotel phone, where the guest is informed that the order is there at the door. The Botlr takes feedback and off it goes.
We even have smart mattresses this is it monitors the guests sleep, automatically monitor the hardness and softness of the mattress where many sensors placed around the mattress capture data while sleeping.
It can also see particular sleep pattern of the guest. It also adjusts the firmness of the matters, position, where the guest is most comfortable.
Using connected devices we also aim to reduce emissions from all our hotels by 30% reduce water consumption by 20% by 2020 . We track water sprinklers, monitor showers, faucets, collect occupancy data, and automatically optimise all these. —Amandeep Sarna, Regional Director-IT, South Asia at Starwood Hotels & Resorts Worldwide, Inc.
Riding a smart scooter
If we were to put our money, we would put our money on clean, connected and smart. It is optimisation, reducing weight of vehicle, reducing energy consumption which makes it an efficient vehicle and clean.
A reliable vehicle is also a cleaner way of building cleaner technology--the overall research utilisation from the time you design, test, prototype, build, sell, own and maintain, till the death of vehicle.
We can achieve this through data which connects the user to the vehicle. If the vehicle knows how fast an 18-year-old boy rides or how long does a 32-year-old woman drive, this is super critical data. As a designer, my biggest fear everyday is fatigue failure, and that is heavily dependent on how a rider really uses my vehicle. So it is critical to know what a vehicle is really going through but today we don’t.
Most people take care of their vehicles but there is a small percentage who might do different things like ride down the stairs or ride really fast.
But with enough monitoring systems and data, they can now design to meet the specific use cases. We should start designing lighter vehicles because today monitoring systems are so cheap. A lot of data analytics is really cheap and free. Today no one understands (battery) cells, people who make cells have a 20% error in their data sheet. But they also have almost 50-60 parameters tested in parallel.
Smart is an over abused term today with everything being marketed as ‘smart’. It is important to educate the rider about what is going on right and wrong in the vehicle to achieve clean again.
Everything in the future will be connected. And the next step to that will be intelligent vehicles. The aggressive index-or how a person rides varies and it’s important that the vehicle learns you as a rider, how many kilometers you vehicle will last based on temperature, tyre pressure, etc.
If the vehicle knows the time you go home and have not charged the vehicle for some time now, the vehicle will try and optimise itself to get you home by adjusting speed and conserving energy. Electric Vehicles gives us a much better platform to make IoT-based systems. Our S340 model has a touch screen dashboard with GSM and it captures data and sends the data to us. This helps in predictive maintenance.
In the future, we could tell you the best way to get home, maybe pothole free way to get home. —Arun Vinayak, chief product officer of Ather Energy.
Unlock a world of Benefits! From insightful newsletters to real-time stock tracking, breaking news and a personalized newsfeed – it's all here, just a click away! Login Now!
