Home > Blogs > Data Science Is Reshaping the Healthcare Industry
There is an ongoing debate that our society is in the midst of the Fourth Industrial Revolution. This revolution is marked by an array of new technologies that are bringing together biological, physical and digital worlds. i-Nside is a global leader in endoscopic technology. Powered by “applied AI systems” and sustainable mobile technology, i-Nside is developing cost-effective medical devices to boost and improve the healthcare industry all over the world. The products built by this company are capable of taking the pictures of the “insides” of the ears of patients to diagnose the irregularities as well as the diseases after a thorough analysis. This analysis-cum-diagnosis is entirely based on the imaging technology embedded in these products.
1) Building custom “visual recognition” capabilities into i-Nside’s endoscopy-products using technologies like Machine Learning and Deep Learning
2) To ensure the visual recognition results obtained by i-Nside products are accurate as this accuracy is directly related to the “life and death” matters.
i-Nside turned to Clarifai, an AI company that is one of the best and leading names in visual recognition technology, for their technological requirements. Using the 100,000 images of the affected areas inside the ears, Clarifai built a custom software for i-Nside products. This software enabled the products to “not only take the pictures of the ear but also to analyse the results – all in one small, affordable, mobile package that anyone can use!”
The team of data scientists at Clarifai coded their software to enable the endoscopy-products of i-Nside to deliver accurate analysis of the affected areas in the ears. Machine Learning technology was used to achieve the “accuracy” part of the whole diagnosis process at i-Nside.
•i-Nside now delivers accurate diagnoses to doctors across the world.
•The machine-learning-based software designed by Clarifai continues to learn and get smarter based on the increasing collection of images by i-Nside-hardware. This is helping the underserved markets like Asia, Africa and South America gain access to the expert knowledge in medical care.
•i-Nside plans on expanding their AI-based imaging hardware into other medical fields like Oncology (cancer) and Radiology (X-rays, MRI scans)
To read more about how i-Nside is disrupting the Healthcare with cutting-edge technology, please follow the link: http://bit.ly/2lLD2ie
Home > Blogs > Tata Sky Training Feb 2017
When we hear about Tata Sky, the word which rings in our ears is Jhingalala. There are more than that.
Tata Sky is the leading direct broadcast satellite television provider (in common words DTH: Direct to home service provider) in India. Formed as a JV between Tata Group and Sky group 15 years back, it has partnered with global leaders in bringing the state of the art digital technology to many our homes.
When we met Tata Sky, about 6 months back we were told that the technology team has to learn about digital marketing so that it can work in tandem with the marketing and communication team. And give the best support while marketing Tata Sky through online media.
At this point all seemed good.
We brainstormed and ideated on a 3 day exhaustive workshop covering the important aspects of digital marketing. The challenge was however to balance between the breadth and depth of the topics to be covered. And yet make it interesting (2 of the training days were planned to be held in weekend) all through out.
Now that was a challenge.
We created a couple of different approach and opened them for feedback internally. We then got it validated with Mr Seshadri (one of the key people in the technology team). He had some insights and some his ideas helped in shaping the final content for the delivery. Now it was ‘adequately’ tweaked to take on the ‘tech’ mindset of the participants.
We started the process by making the participants go through our online digital marketing course. The idea was to get everybody on the same page on understanding the important aspects of digital marketing – mainly the channels and communications. We were surprised to see many of them going on to complete the online course before the workshop. And they were looking forward to these three days. Here is what some of them felt about our online course.
And then …there was the Workshop
Finally on February 2017 when we did the workshop, we started with setting the expectation.
Some of the key expectations that emerged
- Be on the same page with the marketing team (and agency) during campaign planning and relating to the ‘why’ of the marketing plan. This will enable them to add value to the campaigns. Much beyond just giving some technical support.
- To get all those jargons de-mystified and understand the science behind the online marketing execution.
- How to calculate ROI on the money being spend, and how to know whether the desired results were obtained (this is quite surprising from a technical team!)
- How is the competition doing and what can be learnt from them.
After having taken up multiple workshops in various corporates we were ready in our mind to tweak around the content and specifically focus on the areas participants want to learn more on.
It was great to see that at the end of each sub topic there were discussions and debates amongst the participants. Some of them having worked closely with marketing team pitched in with challenges in execution and their view points. And they enjoyed learning about the latest trends like programmatic advertisement and discussing the case studies on related industry or comparable strategy.
It did not end here, they also volunteered to take up post lunch games to drive away the sleepy feeling after some nice lunch. So nice of them!
The results …
(At least as we see it). On the 3rd day we can see the ‘we will now make a difference to the digital marketing in Tata Sky’ look in their eyes. (but they had to wait for the next working day). And the key areas they wanted to contribute more post this workshop were
- Improving and optimising both the websites – the corporate and the self-serve one in areas of user experience, loading time, features etc.
- Look at their web analytics and make data mapping better to improve on the online sales and cart abandonment.
- Helping in creating assets for online advertisement like landing pages for better conversion.
- Implementing some of the new trends and techniques like AMP (Accelerated Mobile Pages)
- Suggest on some of the SEO parameters and directly contribute to the on-page technical optimisation with regular audit in link building.
- And possibly building up ‘branded contents’?
So as the online course and workshop provider we were elated spending three days with eager participants enriching the delivery by their perspectives. We had a few things to learn as well.
This feedback sums it up pretty well.
Home > Blogs > What Powers NETFLIX Homepage?
Netflix Inc. is an American multinational entertainment company offering exclusive movies and TV shows online. The world has gone online and adopted “digital” as the mainstream way of entertainment over that last few years. People are watching TV shows and movies online like never before, Netflix has become a powerhouse of digital online entertainment. With a user base of 57 million, it became imperative for the media giant to offer customized services to each user. But it just sounds easy. In reality, it is an extremely complex task to customize their homepage uniquely to serve content to the users on the lines of user interests and preferences.
- Showing relevant content to the users visiting the Netflix Homepage so that each member finds something interesting to watch – be it TV shows, movies, videos or documentaries.
- Allowing the users to easily explore the catalogue from their Homepage without letting the interface constraints disturb the user-experience.
It is a challenge because there is a variety of content available to the users, thereby spoiling them for choice. But only the relevant content will make sense to users when presented in a coherent and easy-to-navigate manner.
Netflix wanted a solution that would predict what a user intends based on the history of the user-interaction with the homepage via a registered account. Naturally, there are multiple predictions needed to figure out the best content as per the user’s preferences. So, Machine Learning came as the proposed solution to analyse, predict and customize the Homepage content so that users make most of the situation by consuming the relevant as well as interesting video content. The present Netflix Homepage on most devices has an array of two-dimensional rows organized thematically. Users can scroll the video content vertically as well as horizontally.
Various Machine Learning techniques and algorithms helped Netflix personalize and recommend relevant videos to each user. Further, with the help of Machine Learning and predictive analytics, Netflix can predict the ratings a video is likely to receive from the users. Netflix now ranks each video in a row as well creating meaningful groupings of those videos. It is the scoring function of Machine Learning that works at the core of building a customized Homepage for each user.
Netflix has gained huge popularity pertaining to the video recommendations to their users. This helps in keeping the user base loyal and at ease. Not to mention, new users can be acquired quickly too.
Highly efficient and significantly improved online performance across the world.
Users can now find most relevant video content without excessive searching and sifting the Netflix catalogue.
To read in detail about the advantage that Netflix has over its competitors, please follow the link: https://goo.gl/xZeWkq
Home > Blogs > And the Oscar goes to...
La La Land.
Oops ...hold on.
Drip marketing : (borrowing from email marketing terminology). They had a definite plan when they released the teasers, the 1st just hinting it being a musical and revealing as less as possible. And then the second trailer with one of the protagonist singing about 'the mess' they have made (in the movie). And finally the "official trailer" starting with the failed audition of a struggling actor. As we say, that's how you create a hook and connect with your audience in digital media.
Home > Blogs > Data Science-Powered AI Predicted Trump’s Victory
If you’re a Simpsons fan, then you know that they were the first to predict that Mr. Donald Trump would be the President of the United States of America. In the episode ‘Bart to the Future’ (back in 2000) they talk about how Trump’s presidency left the country in financial trouble, apart from other things. If you’re a keen believer in data science and what it can do, then you know there is also a data science-powered AI that predicted Trump’s victory. Unlike the Simpson’s episode, MogIA used 20 million data points to make this prediction.
The system studies data from systems like Google, Twitter, and YouTube. Although Ms. Clinton spent more on ads and set up more field offices, Trump always had more mentions on Social media. His support on Twitter and Facebook was much more than what Hillary had. MogIA was able to accurately analyze this data and predict the right outcome much before anyone else could.
If you look at the number of mentions on social media, you too might think that predicting Trump’s victory is not that tough. But, there’s more to it than that. Not all publicity is good publicity remember? And more interestingly, MogIA can’t analyze if a post is positive or negative.
How MogIA works?
The startup that created this AI tool here in India (Mumbai) in 2004, holds the secret answers. According to MogIA’s developer Sanjiv Rai, it learns in real-time analyzing data on the internet and is not influenced by programmers or developers’ biases. MogIA learns from her environment, develops her own rules at the policy layer and most importantly, it does not ignore ANY data, says Rai.
A lot of people do not want to give a lot of credit to MogIA. Plenty believe that it could just be a fluke.But, this is not the only result that MogIA has got right. Keep in mind that MogIA was so good that it was able to predict the winner BEFORE the FBI took a look at Clinton’s emails.
It’s more reliable as it beats the old-fashioned survey method by miles.
People hate the fact that someone can judge them based on their opinions. So, they rarely share their true opinions. On social media, however, they like, share and post with less fear of judgement. Their data from social networks is more likely to be truer and a more accurate representation of themselves.
Reason for attention
The rapid growth rate of data will create 1.7 megabytes of new information every second for every human being on the planet by the year 2020. Hence, we need means to understand it better.This is why the fields of Data Science, Big Data and Data Analytics are going to get a lot more attention the coming years.
The benefits of being able to understand, organize and analyze this data effectively will have major advantages in the future. If you’re looking for a path with assured growth and demand, you can’t go wrong here.
Leave us a comment and let us know if you believe in the power of Data Science and MogIA or you think that this prediction was just a bit of luck.
Home > Blogs > What is HR Analytics?
Happy to announce the completion of another batch of HR Analytics at the Manipal Global Academy of Data Science in Bangalore. We had 24 participants from top companies.
The idea of the program is to introduce HR Professionals to the applications of Analytics in their work. The workshop is completely interactive with hands-on case studies that participants work through. The cases covered HR datasets which were used in attrition forecasting, employee engagement and recruitment. Participants used analytics techniques like clustering, logistic regression, time series etc. to arrive at decisions.
The feedback from participants is very exciting.
"It was great to learn how HR can enforce its importance and align our measurement metrics to our business objectives."
"Great learning on using tools and stats in best possible manner. Interpreting the results"
"Good Introduction to HR analytics and the possibilities."
"Confident of applying Analytic tools in HR"
"HR relevant case study made lot of sense"
We will be announcing our next batch soon.
Home > Blogs > Data Science Diaries – Rachit Kumar
Senior Data Scientist at LinkedIn
5+ years of experience in the field of data science and machine learning
Previous experience: Affine Analytics
Educational qualifications: B.Tech (Electronics & Communication Engineering)
Specialization areas: Business analytics and Predictive analytics
Rachit recently spoke to Manipal ProLearn about his journey so far as a data scientist. The following are some excerpts from the interview:
Manipal ProLearn: Data science is a relatively new career field. How did you learn about it?
Rachit: Data Science is a relatively new term. Back when I started, it was popularly known as business (or data) analytics. In 2011, right after graduation, I started my job hunt in this space. After a couple of months of rigorous job hunt, I felt that Bangalore is the best place to start in this field as it had a lot of opportunities. Soon after, I landed a job at Affine Analytics.
Manipal ProLearn: Tell us more about your journey so far.
Rachit: Affine Analytics was a start-up. There I got opportunities to work on varying complexities of projects, initially focusing on data preparation for analysis, Business Intelligence reporting etc. As I got exposure to business domains and clients who were industry experts, I naturally developed a business angle to analytics problem solving, which has been of great help.
In the meanwhile, Machine learning was picking up and I started exploring what it really meant. I started taking multiple online courses. Soon, I developed a keen interest in predictive analytics algorithms and the immense, unbounded applications they had. So I started looking for opportunities where I could sharpen my ML axe. Kaggle competitions were my destination in those times, which helped me learn the ML algorithms, optimizing output and their correct execution in great detail.
After a successful and thrilling tenure of over 3.8 years at my first company, I planned to make the next move. I joined LinkedIn as a Data Scientist and started working on some of the key initiatives within the team and the organization.
I have been the lead inventor of two patented (filed) innovations in 1.5 years at LinkedIn and this is just the beginning. Although I have been in this space for 5.3 years, there is always a learning curve. You can never take your eyes off the ball as it can swing away from you anytime and you would never know how and when it happened.
Manipal ProLearn: How will you explain your job role to a non-industry person in one line?
Rachit: I make sure that the quality of LinkedIn’s product metrics is reliable and of the highest standards.
Manipal ProLearn: One book or movie that you think every data scientist should watch/read or could draw some inspiration from.
Rachit: Moneyball (2011).
Manipal ProLearn: What are the three qualities or traits that are a must-have for data science professionals?
Rachit: Keen curiosity to learn about the business domain, statistical knowledge and familiarity with coding.
Manipal ProLearn: What do you enjoy the most about being a data scientist?
Rachit: Understanding the business first hand and gaining insights, which are capable of answering some of the most difficult questions.
Manipal ProLearn: Name any industry leader who has been an inspiration to you, and why?
Rachit: I am a big fan of the product sense and innovation, which Steve Jobs had.
Manipal ProLearn: If not a data scientist, what would be your alternative career path and why?
Rachit: I would be an innovator and leader to create products, which would help everyone achieve their potential. This is also tied to my core belief of empowering everyone to do what they could and would do the best and be proud of their journey.
Manipal ProLearn: What do you like to do in your spare time, when you are not deep diving into the data world?
Rachit: Spending time with family, exploring new mobile phone apps, playing cricket or other sports, watching movies, exploring the internet to learn the incredibly vast number of concepts and things out there.
Manipal ProLearn: Can you share a few activities that can help data science professionals hone their skills? Or something that you like to do to sharpen your analytical skills?
Rachit: Ask yourself periodically:
1. Why do you want to stay in this field?
2. What’s something you definitely want to achieve here?
3. Are you on track to achieve it?
Your answers will help you measure your progress and efforts.
There will be times when you won’t have answers to some or all of these questions. That is mostly a good sign as it means that you have more opportunities than you know. So keep hunting.
Manipal ProLearn: Share the biggest challenge you have faced so far in the field of data science.
Rachit: Most of the big challenges I have faced have revolved around data unavailability and unclean data.
Manipal ProLearn: What do you enjoy the most about being a data science professional?
Rachit: Everyone is interested in knowing what we (data scientists) do and we get a chance to know what they do. I also think it opens up great possibilities for collaboration and innovation, which is the way forward.
Manipal ProLearn: Which was the last book you read/last movie you watched and why did you like/dislike it?
Rachit: I recently watched ‘Dangal’. I liked the incredible zeal displayed by the family in the middle of adverse circumstances. I also liked the script and cinematography.
Manipal ProLearn: If you had the choice to be one fictional character for a day, which one would it be and why?
Rachit: It would be “Shaktimaan”. Even though he is powerful, he is humble and follows his mission every day.
Manipal ProLearn: If you ever have a movie made on your life, whom would you want to play your role and why?
Rachit: I would like to play it myself because I love acting.
Manipal ProLearn: Once superpower you wish you had…
Manipal ProLearn: Describe yourself in three words.
Rachit: Learner, compassionate and interesting.
Manipal ProLearn: What changes do you expect to see in the ever-evolving data science world in the next few years?
Rachit: With the advancements in machine learning, artificial intelligence and machines’ computing power, I can envision a world that would innovate much faster than ever and machines would be increasingly powerful and impactful. It will result in a majority of day-to-day activities becoming completely automated and significantly less error-prone. Like every innovation, it also has the potential of being misused, which could have huge implications on our lives. So the bigger challenge would be to safeguard the world from these severe consequences.
Manipal ProLearn: What is your message to budding data scientists?
Rachit: Whenever you are given a problem, think about all aspects of it. However, never zoom in so much that you miss the bigger picture, which needs to be kept in mind while designing any solution. This could be about the company’s business model, vision/mission or philosophy or even the idea behind a product or its features.
Don’t spend too much time on recurring tasks, take some time to automate them and then monitor them effectively. Use this time to learn more. You will feel better than doing the same task over and over again.Don’t settle until you find what you enjoy working on.
Manipal ProLearn is thankful to Rachit Kumar for taking out his time to talk to us. We wish him all the very best in his future endeavours!
Keep checking this space for more interesting and informative conversations with data science professionals.
Home > Blogs > Fighting the Crime Before it Happens
We have all heard of “aftershocks” of an earthquake. These aftershocks are the smaller tremors that spread to the neighboring regions of the epicenter after largest shock of the earthquake. Well, the same concept holds true for the epicenter of crime. It simply means that the risk of crime increases in the vicinity of the area a crime has already occurred in. So, predicting the possibility of crime and taking precautionary measures can help fight the crime before it occurs. Back in 2002, “Minority Report”, the Hollywood movie” laid the foundations of such a possibility of predicting and preventing crime but the idea was more of a science fiction than the reality. The LAPD (Los Angeles Police Department) turned to Big Data to power their initiative of fighting and preventing crime before it occurs.
Fighting and preventing crime before it occurs.
After the crime has occurred, predicting the patterns of criminal activities in neighboring areas using “aftershock” concepts.
Predicting the likelihood of the next crime is a lot more rewarding than investigating a crime already committed. Damage can be reduced to minimum or zero with accurate predictions of imminent crimes.
LAPD employed a mathematical model, used in predicting aftershocks of an earthquake, to further their “crime-prevention” initiative in the LA region. Developed by Assistant Professor George Moher, this mathematical model is proficient in defining patterns that can help in predicting aftershocks. Predicting an actual crime is very difficult but predicting the aftershocks of a crime is relatively easier. It is so because there is always an increased risk of a crime taking place in the vicinity of a bigger crime that has already occurred. This fact is proven and there is enough data to substantiate the fact in terms of historic data on crimes. Mostly, crime data shows similar patterns. So, LAPD fed the mathematical model with data for 13 million crimes committed over last 80 years. Based on this data, the model generated predictions (aftershocks) that matched with the actual crimes that had taken place in the vicinity of where the bigger crimes were committed. Using Big Data, the LAPD improved the prediction algorithm. Now they use the various data sets related to known crime hotspots to predict the likely places where crime can take place. The LAPD places officers in the predicted locations at the right time so that they can prevent the crime from happening.
- As an extended service of this LAPD counter-crime initiative, insurance frauds are also detected and curbed using the Big Data predictive analytics.
- Apart from preventing crime, this initiative is helping in understanding the “intent” and “source” of crimes. With this understanding, criminal psychology will be studied and predicted more accurately.
- The LAPD recorded a 33% reduction in burglaries, 21% reduction in violent crimes and 12% reduction in property crimes after using the predictive algorithms.
Home > Blogs > Valentine's Day through the eyes of a Data Scientist!
Happy Valentine's Day everyone ! While we celebrate the day with our loved ones, Team ProLearn decided to look at this day through the eyes of a Data Scientist.
Did you know that the total worth of merchandise bought on Valentine's day is more than Rs 5 Billion or that more than 20 million roses are sold during Valentine's week? Well, Data Scientists would. They would also know that in the week leading to Valentine's day florists are expected to notch up sales of Rs 250 Million plus or that greeting card is the single most profitable item during Valentine's Day and sales of cards/gift items rise by up to 10 times just ahead of 14th Feb. Data Scientists are able to mine user data, purchase data and sales trends to arrive at such insights and this is what makes them so valuable for businesses. After all which retailer would want to stock up on shirts when data says the most sought after Valentine's day gift for men is the humble silk tie.
No wonder Harvard Business Review named Data Science as the hottest job of the 21st Century.
Home > Blogs > How training Employees on MS Excel helped Faiveley Transport predict future production & optimize output
BACKGROUND AND COMPANY
Faiveley Transport is an international innovator, manufacturer and supplier of equipments for the railway industry. Established in 1919 in France, it has now firmly cemented its position as a leading company in its respective sector. The company’s scope of supply includes the study, design, manufacture and the delivery of the required equipment. The company is also a major player in the railway sector of India, and has been a major equipment supplier for Delhi, Bangalore and Hyderabad metro systems.
In 2016, Faiveley Transport extended their capacity and capabilities at Hosur site, near Bengaluru, in order to cover the entire range of their product portfolio. The site thus grew, witnessing an increase in activity and turnover. However as a result of the expansion, the volume and complexity of their operations increased, which necessitated training of employees in a comprehensive MS Excel database reporting application. Such a training was also important to streamline internal business processes, workflows and communications.
OUR ROLE AND SOLUTION
It is at this juncture that Manipal Prolearn was roped in by Faiveley Transport to provide advanced and specialised training to the company’s employees in MS Excel. Although everyone is well versed with the basic functioning and use of MS Excel, most professionals require some form of training to grasp the more complex functions of MS Excel tools. The Advanced MS Excel Online Certification program is thus designed for an entry level professional or a senior level manager who can learn advanced MS Excel skills applicable to a wide range of industries.
BENEFITS TO THE CLIENT
In addition to providing with an online certified course, Manipal Prolearn supplemented the online training with a 2 day workshop for the employees of Faiveley Transport at the Hosur site. The workshop enabled the employees to put to use their online training to fit the unique needs of their company. Post the training and workshop, Faiveley Transport employees have mastered MS Excel tools such as formulas and macros to predict future production and optimize their output, create reports based on production data and make presentations and marketing pitches. A significant amount of data has been collated in one comprehensive database, allowing the company to visualise and analyse the data, and thus becoming more productive and efficient.
As part of the offering, the L&D team at Faiveley also received weekly dashboards with regards to learner progress, faculty support for query handling and resolution and a state of the art learning management system that enabled learning on the go.
ProLearn is privileged to have made an impact in capability development at Faiveley and we have now embarked on a journey of streamlining the project management needs of the client.