VGVinay Gupta: We are in Renewable Energy generation sector, so maximum uptime of Wind turbines and generation of maximum power are the main focus of our operations. Sensor analytics carried out by us is aimed at maximizing the revenue for the customer through enhanced performance, availability and reliability of Wind turbines. In addition, we also aim at reducing the operational costs by continual process improvement. We have an end to end Enterprise Analytics Eco- system viz. Data Warehousing, Analytics Application and Real Time Visualizations.
In our organization, we continuously strive to enhance the quality of data being generated/ captured so as to ensure higher availability of Sensor data, reduce rework on data generated before analysis and build confidence across the users for Decision Making at Strategic, Operational and Tactical Level. A lot of steps have been taken in this regard by re-engineering of communication systems, process improvements and employees training.
Further, we spend considerable time in formulating the right question, because then time and efforts spent in analytics would yield the maximum dividend. We firmly believe that data driven facts should drive our major decision making. The recommendations/inferences obtained from data analytics are discussed, brainstormed and complemented with past experiences and inherent domain knowledge, before the decisions are taken.
AIM: What is your approach to face the challenge of meeting needs of so many clients across vast geographies with limited resources?
VG: I feel that the solution lies in the basic heart of Analytics process and eco-system itself. To meet the challenge of large number of clients, which in our case includes both internal and external customer, we lay a great emphasis on correct understanding, documenting the need/problem and prioritizing each need/requirement.
Consequently, we work diligently on the algorithms and methodology of analyzing data for those requirements which have the maximum impact on the business. Once proof tested and outputs confirmed, we automate the analytics end-to-end, ie from the process of data extraction to visualization, and embed this into our current work stream. This ensures timely availability of information, greater value to the user and low overheads on a continuous basis.
AIM: What are the key differentiators in your analytical solutions?
VG: In energy sector the key data sources are quite different from other popular sectors viz. Retail, Banking, Social Media, Hospitals etc. We have a large network of sensors which are embedded in the wind turbines, which generate huge volume of data in a continuous manner. These data packets are transmitted in real time to our central storage system. Thus there is a need to analyse this data on a 24 x7 real time basis to understand the performance of each wind turbine.
Based on certain alerts, decisions need to be taken in a few seconds. In addition, the performance of wind turbines is analyzed based on various systems/sub systems and material/spare parts used. Thus the key differentiator is the integration of sensor analytics, engineering analytics and materials analytics under one roof. The models deployed are both descriptive and predictive ones. We need to migrate towards the prescriptive models, as well.
AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth.
VG: We have a team of around 40 analysts in the Head office and around 30 analysts in seven states in India, where our wind farms are located. We have organized the Business Analytics & Business Excellence Department into four main groups
Transactional Data Warehousing team, which includes Turbine Sensor data and ERP data, Data Modeling and Analytical team, which develops new application and analytical models
Data Presentation team, which includes visualization, dashboard generation and reports creation for internal and external customers
Operational Excellence Team, which takes on business critical projects for end-to-end analysis and solution generation.
We ensure that the skills sets of our team are constantly upgraded and in synch with the current technologies and methodologies in the market.
AIM: What are the next steps/road ahead for analytics at your organization?
VG: In the immediate future, our focus is mainly to consolidate the initiatives undertaken by us to enhance the analytics capability. The next big step would be to identify the right big data technology and tools, which will meet our business requirement. There is a plethora of technologies, tools, data platforms and data integration software in the market.
But making the right choice, which will be a value for money in the long run, is the main challenge. It’s important to have the correct use case for such implementation and one should not do it just because it is in vogue. Another focus area would be to develop applications and expertise in prescriptive analytics and mobile computing.
AIM: What are a few things that organizations should be doing with their analytics efforts that most don’t do today?
VG: As you are aware that there are two sets of complementary activities in the field of Data Analytics. Firstly, Analytics can be considered either a science or an art and secondly, Decision Making is either intuitive or fact based. I feel that the process of using data analytics in decision making as a competitive advantage is a combination of the both the aspects. Therefore the users at different level need to be aware and proficient in these aspects.
Presently, a large number of companies are in the process of collecting/ capturing huge amount of data, considering that would be required for analysis at the later stage. Very little efforts are being put to ensure the veracity and value of data captured at the transactional stage (or Point of Origin) itself. Emphasis in this direction will not only increase the confidence value in the data sets but also significantly reduce the efforts for analytics/visualization and enhance the quality of decision making. Another important aspect is to avoid making incorrect assumptions and failing to test them while analyzing.
AIM: What are the most significant challenges you face being in the forefront of analytics space?
VG: In my opinion, the biggest challenge is to embed the analytics process and insights obtained into the decision making at junior, middle and senior level. Building the culture of data driven approach needs few good success stories. Fortunately, we have been able to make the initial impact in our organization.
The second challenge is to ask the right question or generate the right use cases for business functions. Another, significant challenge is to keep updated with the new technologies being developed and evaluate for absorption in the organization.
AIM: How did you start your career in analytics?
VG: My journey in analytics space began around seven years back in New Delhi, when I started handling huge volume of data related to the multi-disciplinary defence equipment and technical manpower in order to ensure proper planning, employment and resource optimization. Thereafter, I had the privilege of establishing the first Center for Data Analytics and Optimization in Secunderabad. We conceptualized, built and established the complete analytics ecosystem from the scratch.
My present role encompasses Sensor based data systems (SCADA), transactional data systems (ERP), Data Integration, Modeling and Application development, Real time analytics and visualization, Customer Portal and Process Improvement projects. We have built various analytical models for optimization of Wind energy operations to drive the business strategy. Modeling includes predictive modeling, O&M revenue modeling plus forecasting, reliability analysis and BI/BW implementation. We aim to optimize the Wind farm performance through Analysis and Visualization of sensor data (high volume & velocity) of 5500 plus Wind turbines and ERP transactional data on real time basis.
AIM: What do you suggest to new graduates aspiring to get into analytics space?
VG: My advice to the new grads is that they need to understand both business process and analytics methodology for achieving the positive results. With mushrooming of various analytical applications, the complexity of calculation is hidden, but to correctly interpret the result of analysis and provide solutions for critical business problem is the key to success. Business problem- Analytics Problem- Analytics Solution- Business Solution; this chain needs to correctly linked and implemented for the desired outcome.
AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics?
VG: We recruit new graduates as well as experienced individuals, who have an ability to understand the business problem quickly. These professionals should have the passion to play with data, look through the data and generate insights using various statistical modeling techniques, both descriptive and predictive ones.
We look for the skills of understanding and applying various data modeling and integration techniques required in engineering analytics and are not driven by a particular Analytics Platform or product. We believe that the platform or application related skills can be acquired with short capsules, hand holding and team support. Finally, it is the attitude towards problem solving and team co-operation, which matters the most.
AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe?
VG: The concept and terms for data analysis to gain insights in business process has evolved from Decision Support System in 1970-80’s to OLAP’s in 1990-2000 to Business Analytics & Big Data in 2010 onwards. There are two main sources of Data Generation viz. Firstly, the customer or social media activist (in BFSI, Retail, Social Media, Telecom, Media, Medical etc) and Secondly, the machine sensor data ( Oil & Gas, Energy, Manufacturing, Automobiles etc). Presently, both these sources generate huge amount of data and analysts are trying to understand either the customer propensity or the machine behavior/performance.
The future lies in the integration, where-in sensors would be embedded on/inside the human beings for health, fitness, productivity, location or behavior analysis. It would also enable the sensors to analyse our moods and brain waves and send the information on real time basis to health agency or the sales department. While this concept would further impact the privacy of an individual, but it is considered to be potentially revolutionary in terms of both volumes and benefits.
As regarding the Engineering Analytics (EA), the study carried out by research firm Zinnov Management Consulting in Aug 2013 has revealed that the market is poised to touch USD 27 billion by 2017. The industries like energy, automobiles, aerospace and healthcare will lead this development. Two years before, in 2012, the spending on engineering analytics, which refers to derivation of meaningful insights by processing information from physical machines, was around USD 12.6 billion. So, in five years this market volume is anticipated to become more than double.
AIM: Anything else you wish to add?
VG: To achieve ‘Nirvana’ in Analytics, one need to continuously adapt, change and evolve, especially because the data and processing technologies itself are rapidly changing with time and varying circumstances. I feel that even though there are a large number of machines operating with sensors and performing business critical applications, and large amount of data is available from them, Engineering Analytics will still take some more time to rise to the center stage of the Analytics world. Machine behavior and performance analysis could raise the efficiency levels beyond what has been achieved till date, and therefore need to be given due significance by users, engineers, analysts and statisticians of the future.