Predictive analytics - one dimension to tap max potential of your Big Data

Predictive analytics is the process of using data analytics to make predictions based on data. This process uses data along with analysis, statistics, and machine learning techniques to create a predictive model for forecasting future events. When we say “predictive analytics” we mean   application of a statistical or machine learning technique to create a quantitative prediction about the future.
For example False Alarms from electrocardiographs and other patient monitoring devices are a serious problem in intensive care units (ICUs). Noise from false alarms disturbs patients’ sleep, and frequent false alarms desensitize clinical staff to genuine warnings. Metlab has developed an algorithms that can detect QRS complexes, distinguish between normal and ventricular heartbeats, and filter out false QRS complexes caused by cardiac pacemaker stimuli. The algorithms produced a true positive rate (TPR) and true negative rate (TNR) of 92% and 88%, respectively. What it means is data of ECG, ABP & PPG devices are now being processed to identify a false alarm and it is improving life of patients as well as hospital staff.


Predictive analytic's has become the key to helping businesses create differentiated, personalized customer experiences and help customer make better decisions. As we know predictive analytics strategy and architecture are very custom to the client landscape & requirements.

Key considerations while architecting for Predictive Analytics-

1. Predictive analytics must cover the full customer life cycle                                                       . Organizations require predictable insights into customer behaviors and business operations. Design idea should be to deliver value to customers throughout their life cycle to differentiate their customer experience and sustain business growth. The business stakeholders input is key and it will help identify effective mechanisms for translating the business knowledge to predictive algorithm inputs thereby optimizing predictive models faster and realizing deeper customer insights.
2. Build your Big Data architecture around predictive analytic's                                                    There is no dependency of predictive analytic's on big data, Only thing that has changed is with Big Data processing e have access to larger data sets that can be used for predictive analytic's.  By putting predictive analytics solutions at the core of your platform we create synergies between the analytic's layer and big data processing.
3. Learn from leading companies embracing digital transformation.                                         A large Australian telco improved customer discovery for precise targeting of ads; Another US telco optimized its marketing mix for effective channel interaction; A UK telco major deepened its insights into the voice of the customer to improve customer satisfaction; and many companies are providing personalized digital experiences to maximize customer loyalty.

Career choice - What has better future Android Development or Big Data Development?

I was at a Big Data seminar and one of the participants asked me this questions about choosing career between Android & Big Data. Incidentally I have experience of working on both technologies as an architect/developer/designer and I found the dilemma interesting for someone who has to choose a career track and I decided to post my thoughts.

Choosing a career track is a difficult decision when you are a fresher and don't have any development experience. My suggestion is to first spend some time& effort in learning about both these technologies. If possible you should attend a basic training and do some hands on development on both technologies to get an idea of what is it all about.

Both technologies 'Android' and 'Big Data' (Big Data is actually a set of technologies) have tremendous potential and while Android is already accepted and adapted by industries, Big Data is still in the stage where more than 80% of the industry is still planning their move. Architecturally Big Data is more complex as it is about restructuring the approach to process and consume your data and requires experienced data architects to plan the Big data journey where as android /mobile architecture is kind of 'add-on' that can sit on your existing architecture and does not involve huge invests. So investment is also a decision factor for companies to adopt Big Data thought you can always start small with Big Data. Big Data is actually a set of technologies from Haddop, HBase, Hive, MAP-R  to Data Analytics and more. In the long run there is an opportunity to become a data scientist and we are already short of data scientist and the demand for data scientist is expected to grow exponentially. So is Big Data a good career track ? Definitely it is a interesting and challenging career track and this is a good time to get started on Big Data set of technologies. Do

The reason Android (or mobile architecture in particular) is a hot technology (and has been so for last few years) is because the exponential rise is use of mobile devices and improved & cheaper mobile internet connectivity across the world. Today the companies do not have an option but to make their services available over mobiles for their consumers as well as their work force. Android market share is increasingly growing as compared to iOS so it is a definitely good career track to aspire for. The other reason why one can choose mobile technology as a career track because there is big potential innovate in this area and you can see many startups leveraging mobile technology and it does not require huge investment. One good idea & an innovative mobile application can change the game for you. I am fascinated by the potential of these technologies and have alternated between the two (fortunately). So get trained on both the technologies and then make your move. Best wishes!
   

Blogmint If your blog is influencing people, why not let it influence your bank account?

I recently went through the Blogmint.com. Blogmint.com has a unique model that serves brands and bloggers.Bloggers have reach in the social media and their own nice followers based on the content of their blogs. The brands want to reach-out to the consumers using trusted channels and Blogmint offers the platform. Various brands register and launch campaign programs and registered bloggers can show interest in a brand campaign they are familiar with and if the brand accepts their blogger credentials the blogger can blog/write about their experience with the brand and get paid for their writing. Interesting idea because the blogger can 'Choose a  brand he is familiar with'to blog and he is not forced to write for 'any brand'
Here is the link to Blogmint.com in case you are interested in knowing more.
Link : http://blogmint.ref-r.com/c/i/7230/3008573

Downgrading from Windows 10 to Windows 8 when your laptop manufacturer does not support Windows 10

Downgrading from Windows 10 to Windows 8  ( when your laptop manufacturer does not support Windows 10 or if your system becomes sluggish)

Recently I started facing issue with my Dell 14Z Inspiron Ultrabook. I had upgraded OS from Windows 8 to Windows 10 around 2 months back. After upgrading to Windows 10 I felt the system became slow and started giving various random errors. When I checked with Dell Support they advised me to revert to the OS factory version which was Windows 8 in my case. Thankfully as soon as I bought the laptop I had created DVD backup media to restore the system to factory image in case of failure ( I had created a 4 DVD set using Dell Backup & Recovery software).

I took backup of all my data and software and then I tried to use the DVD media to restore to factory image but there was some issue and the backup would not run, it started giving me Boot error (Error Code- 0xc00000d ).

 After lots of troubleshooting I could not fix this error and then reached out to Dell Support and they told me they can send me a USB with Factory Restore media for my laptop ( Fantastic Support! If nothing else one should buy dell laptops for their after sales support. IN my case I did not even have a extended warranty yet they provided dedicated support and even send the USB media without charging me anything) .
Once I received the USB with Dell Backup and Restore media I changed the setting in laptop bios to boot from USB instead of the default boot from disk. The restore failed for the first time because the default setting on my laptop was 'Secured Boot' as pointed by very helpful Dell support. Once I disabled the Secured Boot in my bios I was able to complete the factory restore to Windows 8

After the install I checked the driver and I found couple of drivers needed update (See image below)

Once the drivers were updated the next step was to check for Windows updates. There were around 190+ Windows updates that needed to be installed and once the task was completed the system was completely restored. It has been 2 weeks since I have rolled back my Dell 14Z Ultrabook from Windows 10 to Windows 8 and the laptop is faster and stable. Yesterday I started getting messages to update my OS to Windows 8.1. I did not want to upgrade to 8.1 as Dell has not tested Windows 8.1 on my laptop model ( Frankly there is no harm in upgrading from Windows 8 to Windows 8.1 as there are few changes and all these changes are to resolve some cosmetic issues present in Windows 8 but my recommendation is check with laptop manufacturer before upgrading Windows OS).

How to stop Windows from forcing you to upgrade to Windows 8.1

Check this useful link to disable automatic upgrade to Windows 8.1
: http://www.askvg.com/how-to-stop-automatic-forced-upgrade-from-windows-8-to-windows-8-1/

I hope this post helps those wop want to rollback from Windows 10 to Windows 8 after the 1 month rollback period is over. I struggled a lot initially but with help of Dell support it was easy to roll back to Windows 8. If you are a Windows user, one thing to note is that if laptop manufacturer does not support Windows 10 then you may think of changing your laptop in near future because Windows will soon stop support for Windows 8 and then you will be stuck with your old laptop. You can of course put the old laptop to some good use. I have a Samsung X05 Pentium Windows XP laptop with 30GB HDD. There is not much you can do with a laptop that has 30 GB HDD but if the screen resolution is good then you can use it for surfing and watching videos. I have put the XP laptop in my kitchen to view cookery videos while I am experimenting with cooking. Good luck!

How to use Gapminder DatatSet - R Language

Here are some quick points I picked while experimenting with gapminder dataset
Before you can work with the gapminder dataset, you'll need to load two R packages that contain the tools for working with it, then display the gapminder dataset so that you can see what it contains.

Note - Install  gapminder & dplyr packages before trying following commands. You can search this blog for instructions on how to install R packages.

##### Code Sample #######

##### Load packages gapminder & dplyr

library(gapminder)
library(dplyr)


# Filter the gapminder dataset for the year 1957
gapminder %>% filter(year == 1957) 



# Filter for India in 2002
gapminder %>% filter(country == "India" , year == 2002)


---

Flooding in metro cities of Mumbai & Chennai. Act now to avoid future flooding or will we drown in the next rains

The administration in Mumbai, Chennai and other major cities is ignoring the basic issues that are leading to flooding of the cities. The civic officials are too busy to divert attention rather than highlighting the efforts to prevent flooding.

1) Use of plastic bags - On an average every citizen shops twice a week which means every citizen contributes 8 plastic bags to the city waste every month. So 22 million population of Mumbai adds at least 176 million plastic bags every month! Apart from plastic carry bags there are other types of plastic that gets thrown around the streets by people for example plastic packet of wafers, chocolates, biscuits etc. So my estimate is Mumbai is contributing close to 250 million plastic bags/packets every month. People do not realize that 250 million bags are being added every month - they are accumulating and the soil is loosing the ability to absorb water. So there is no time left but to act right now or we will have worse flooding next year that will kill hundreds of people.

Solution options - There should be absolute ban on plastic bags and anyone carrying a plastic bag should be fined on spot. Unfortunately there is no other option but to be harsh and strict to ensure no more bags are getting added to existing problem that we have.

2) Not Recycling plastic bags - Are we recycling 250 million plastic bags every month? To add to the problem Mumbai city does not have space to dispose the garbage. Civic administration digs pits and dump garbage in the pit for most of the garbage except for the few modern garbage disposal plants, that do not have capacity to handle the entire garbage produced by the city. So most of the  plastic bags are not recycled and end up in drainage systems that clog the drainage pipelines.

Solution options - We need private companies to help setup garbage treatment plants particularly to treat plastic and non biodegradable garbage. Government should also setup 'Plastic collection points where citizen should be encouraged to deposit their plastic bags and plastic items so that they are recycled and do not end up blocking our drains and blocking oil from absorbing water.

3) Concretization & Tar-ification of soil - Modern Smart Cities have reduced the soil cover. If you compare the satellite picture of Mumbai over 5 years we will see that the soil that used to absorb rain water has now been covered by cement an tar. What this means is the rain water is not getting absorbed by the soil and the only way for the water to escape is by drainage system which were not designed to carry huge volume of water.

Solution options - The builders and residents should realize that thy have to play an important role in maintaining the environment. If we cover the entire building premise with cement and roads with tar then we are blocking the rain water from seeping in and inviting floods. Do not use concrete to cover land inside your buildings and parking lots. 

Indian banks are not leveraging digital technologies as they should

I have bank accounts in couple of leading Indian banks for over 15 years and till date none of these banks have tried to cross sell a single product to me.
1) They have never inquired or tried to sell me an insurance
2) They have never given me advice on how to invest money lying in my saving account
3) They have never tried to sell me other banking products like MIP, FD, car & Personal loans

In short the bank are spending big money to increase their customer base but they not utilizing their existing customer data to cross sell their other products to their loyal customer. Is that because banks are so focused on locating the forest that they are missing the trees? Bankers expert in core business so not have access to smart technology brains who can help them implement cross channel sales? Or are they so bogged down managing existing set of technology systems that they do not dare to venture into developing platforms to enable cross selling?

To give an example, when I wanted a Unit Linked Insurance Product I had to do market research and finally I ended buying a product from the bank that I bank with. I was surprised that the bank that sends 10s of mails a day did not think that 'Their Privilege Saving Bank Customer' would also need insurance products or other saving products that are offered by the bank. They did not consider that if they do not cross sell than they will lose the customer to their competition? May be the Indian banks are happy with the money they are making from my saving account or maybe they are just lazy to analyse my credit history to find if I have any money lying around that can be put to good use or maybe they don't care about growing my money (which they should).

Banks & their associates should understand that there are 3 negatives of not doing customer analysis

1) Bank or Insurance companies are not selling products to the customer who already have faith in their company
2) It is poor customer service if the bank does not reach out to the consumer to educate and guide them in investing their money and their is added risk of losing customer to another bank that provides such services
3) The bank is spending tons of money to find new customer and set new branches to increase their business and customer base but not tapping the existing customer base which is poor business strategy.

I did not get a chance to work with Indian bank in last 14 years but I would like to sit with the CEO's of these banks and understand why they are not using the customer data in a smarter way and to correct the 3 negatives I have mentioned above. Cross selling is a win-win for all parties and it can give birth to some innovative products and schemes. For example I have often wondered if I had bought an insurance product from my bank would they offer me, 'their loyal customer of 15 years' some benefits or free-bees like free accident rider or critical illness rider or maybe lower premium as compared to non-banking customer? Wishful thinking, Bottom line when banks use their data smartly everyone will benefit  - the banks, their insurance wings, their investment wings, their associate companies as well as their customers. Let me tweet Chanda Kochhar & Aditya Puri and may be they will offer me a position in their strategy team!

Did you know that your car dealers are using Big Data to sell you a car?

Today customers interact with a brand through a variety of channels - from mobile, e-commerce, social media, digital ads to email campaigns. By analyzing customer interaction data across all channels company can turns leads into customers. Big data gives you the insights that are key to optimizing campaigns and targeting customers.
                                                                                      A potential car buyer does web search , gets quotes from websites and reads reviews on internet before buying a car and many auto companies have started using this data to identify their potential customers. You should not be surprised to get a call from a car dealer who knows more about the car you own then you would have remembered.
               Dealers are already designing strategies to target customers from the big-data that is available to them. So before you start thinking of selling the old car you get a call from dealer who explains why you need to change your car and what new features of the latest car  model are missing in your car. Dealers know service history of your 1st car , they know what was the last time you services the car and they also know what accessories you have bought for your car in last 5/10 years, they know how much money you have spent on your car since you bought it and finally they know or shall we say they can predict the next car you are likely to buy.
The auto dealers & insurance companies may not be using Hadoop yet but they sure know how to use the bigger data to sell more cars

Smart Tracking Dashboard - Using technology to monitor, trace and predict terrorist activities

While we mourn the tragic terrorist attack in Paris I wonder what technologies are used by governments and intelligence agencies to guard the country against anti-social and terrorist elements. We wonder what kind of intelligence information is shared by security agencies within the country & with with friendly countries? We wonder if GPS and Mobile CDR data is used to monitor suspicious individuals and known links and also to identify the localities frequented by criminal elements. Terrorist do not appear out of thin air, they live in apartments, they frequent the target areas to plan attacks and they carry cellphones.  By use of software and intelligence it should be easy to predict and prevent attacks like Mumbai & Paris.

One idea is that security agencies of each city can create a 'Smart Tracking Dashboard' that shows all new mobile numbers that 'ENTER' the city on given day (This data can be filtered from the Call Data Record).  New mobile or IMEI numbers would indicate people who are outsiders and monitoring movement of such mobile near high security area & sensitive locations would help security agencies identify any pattern which would help them focus on certain individuals who may have to be investigated or questioned in order to prevent any incident. To give an example David Hadley was carrying a US mobile and did a recce of all areas that were later targeted by the 26/11 terrorist. A proactive monitoring would have helped the police identify a pattern that "tourist  A has been visiting some sensitive locations multiple time over a period of days" which is unlike a normal tourist who would only visit the location once. This is just a simple example but there are multiple events that can be built into the software system by defining a set of events that need to be monitored. What we get is a Smart Tracking Dashboard.
 
The next important thing apart from building Smart Dashboards is to built a system that enables sharing intelligence across Departments & agencies in Real Time. This will remains to be the critical aspect of predicting and preventing threat elements from causing harm to society. Wider intelligence data will enable better predictions of threats within the country and across friendly countries. 

Key takeways for our security agencies

1) If there are people who are suspected of terrorist activities or sympathizers of certain terror groups then security agencies will have to use IMEI and mobile data to monitor their movement and activities using sophisticated , smart and self learning software 

2) Today no country has an option of not monitoring terror suspects for what ever reasons and lawys have to be modified to allow surveillance  

3) Global terrorism mandates intelligence sharing across countries. If a 'suspected or known terrorist' leaves country A to visit country B, then there should be proactive notifications sent to country B so that they can be monitored, Many incidents cab be prevented if we have smarter way to monitor the city.

How to SPEEDUP your Android Studio ?

Here are some quick trips that I have tried to speed up Android Studio on my Windows 8.1 Laptop. I hope you have a laptop with 4GB RAM because that is recommended.

1) Un-Comment  (if there are commented) or if they are missing then add these 2 lines to the project gradle.properties file (See image below)

org.gradle.parallel=trueorg.gradle.configureondemand=false 

2) Add these 2 line to the gradle properties file at the location - C:\Users\DataScience\.gradle where 'DataScience' is the user that I have logged in with to my Windows machine. If your Windows user is Jim then the path for you should be C:\Users\Jim\.gradle

3) In android studio, you can enable offline work to make android studio run faster. To enable offline in android studio, just follow following step.  In Android Studio, go to file >> setting and click Compiler from side menu and then type --offline in the command-line options box and hit OK button like this: