12. 12/11/2016 12
EVOLUTION OF DATA
0
5
10
15
20
25
30
35
40
45
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Data (ZetaBytes)
Next Generation BI Analytics
Source: IDC
13. EVOLUTION OF DATA MODELLING
12/11/2016 13Next Generation BI Analytics
14. EVOLUTION OF DATA VISUALIZATION
12/11/2016 14Next Generation BI Analytics
15. EVOLUTION OF DATA VISUALIZATION
12/11/2016 15Next Generation BI Analytics
16. 12/11/2016 Next Generation BI Analytics 16
DELIVERY APPROACHES
FOR
ENTERPRISE BI ANALYTICS
Corporate BI
IT Managed - Self
Service
Business Led
Utilization of reports and dashboards published by
IT for business users to consume
A “managed” approach wherein reporting utilizes
only predefined/governed data sources
Utilization of reports and dashboards published by IT for
business users to consume
17. DELIVERY APPROACHES
FOR
ENTERPRISE BI ANALYTICS
12/11/2016 Next Generation BI Analytics 17
Category Approach Managed by Ownership Scope for
business users
Governance
Business led Self
Service BI
Bottom – Up Business All elements are
supported by
business
Data
preparation, data
modeling, report
creation &
execution
Business
IT Managed Self
Service BI
Blended
Approach
IT IT : Data+
semantic layer
Business :
Reports
Creation of
reports and
dashboards
IT: data +
semantic layer
Business: reports
Corporate BI Top – Down IT IT Execution of
Reports
IT
18. 12/11/2016 Next Generation BI Analytics 18
ALIGNING WITH AGILE MANIFESTO
Embrace change - Reporting requirements are dynamic, tools should be simple for end
user to develop
Deliver working reports regularly – All three types of delivery models should be aligned so
that new and meaningful insights are generated regularly
Strive for iterations – The process from source data to visualization should be iterative.
Test throughout the lifecycle – Testing becomes very critical when business does their own
reporting. IT should pass testing practices to business and they should run through out the
process from data to visualization
Getting the architecture correct by experimenting all suitable options through POC
21. 4 4
4
3 5
5 4
3
5
4 4
4
3 4
3
5
3
4
3
3
5
4
IMPACT ON CLIENT: Client can reach to many subscribers by sending limited number of campaigns
12/11/2016 21Next Generation BI Analytics
SOCIAL NETWORKING ANALYTICS
Social networking analysis can be performed based-on degree of centrality
SNA using Degree of Centrality
Step 1: Find out the number of direct ties to the other people in
the network for subscribers with high churn probability
Step 2: In order to spread retention campaign to as many
subscribers, we will find out subscribers having higher degree of
centrality. Assuming, if a subscriber with high degree of
centrality is churned then probability of its peer group getting
churn is high
Step 3: Send campaigns to the target subscribers (found in
step 2)
Degree of Centrality
IT Managed Self Service BI Example - demo
22. NEXT GENERATION- BI ANALYTICS
12/11/2016 22Next Generation BI Analytics
Corporate BI Example - demo
23. 12/11/2016 23Next Generation BI Analytics
DATA IS LIKE A BIKINI!
What it reveal is suggestive, but what it conceals is vital.
"Analytics can reveal everything" Do It Yourself (DIY)
Good afternoon . Hope you had a good lunch. So can I have a loud Good afternoon. Would you like to have another one
I am here to demonstrate how to prepare Konda Kavum.
So we will take a bowl of rice flour, a bowl of wheat flour, water and brown sugar and mix it.
So here is the Konda kavum. So I am not going to fry it but give you as it is. So who like to have.
Why don’t I offer you something prepared by professional chef with evenly cooked and presented in a nice manner.
How many of you would like to have the cooked one instead of the earlier one
So what impacts your decision. It is the presentation or the process or both.
So my topic for today which is analytics is similar to cooking.
Good afternoon . Hope you had a good lunch. So can I have a loud Good afternoon. Would you like to have another one
I am here to demonstrate how to prepare Konda Kavum.
So we will take a bowl of rice flour, a bowl of wheat flour, water and brown sugar and mix it.
So here is the Konda kavum. So I am not going to fry it but give you as it is. So who like to have.
Why don’t I offer you something prepared by professional chef with evenly cooked and presented in a nice manner.
How many of you would like to have the cooked one instead of the earlier one
So what impacts your decision. It is the presentation or the process or both.
So my topic for today which is analytics is similar to cooking.
Analytics is about helping taking a better decision. So I prepared this dashboard for a hospital head to find patient satisfaction. Similar to my dish I put a lot of effort to prepare this dashboard. I have taken a lot of data of places and combined them to come to this visual. Doesn’t it look wow?
Do you think it would be helpful in taking a good decision
Is it a self explanatory dashboard
So I just want to emphasize the importance of Data Visualization is taking a better decision.
Let me take you through a step by step process for Analytics
Data is most important part of taking any decision. If data is not correct then even the best modeler cannot help you in creating a good Dashboard
If modeling is not done the proper way the outcome may not be relevant.
Presentation of any dashboard is vital to take any decision. If your dashboard is not simple and not fulfilling audience requirements then the whole effort to prepare the one is gone in vain.
Taking the right decision is definitely important and should be based on the process rather than the gut feeling.
Forecasting based on historical data is a crucial part of analytics and now a days we are actually using a lot of statistical techniques to forecast better.
Why are we talking about analytics now a days. Because Analytics is top priority of CIOs now a days. According to IDC data would be 40 zetabytes by 2020.
Data is unstructured rather than structured.
The Mico SD cards storage is increasing as so in Hard disk
Big data includes CRM, ERP, WEB, GPS etc.
Who does Data modelling
Earlier only Data modelers were doing it but now with the advent of self service BI tools like PowerBI it is being done by end user.
However as our wife/ mom cook food in home but when we need to eat something special we go to a restaurant to get it prepared from professional chef so when there are complexities then experts like us come into picture.
Earlier the end user used to give the requirement and it role was to prepare the report, govern the data etc. But now a days using self service BI tools business user create their own reports and IT role is limited to governance and enabling business users to create their own dashbaords
Integration of Sources
Creating Dashboards
Forecasting
Sharing
Q&A
Lets see what is this. So everyone of use have friends on Social Media some have more and some less. Lets say I have 50 friends and you have 10 friends so my Degree of centrality of circle of influence is more than you. The assumption we have taken here is that no of connections = no of influencers. So in order to make the campaign effective we need to target people having more friends. So if we are able to retain him his friends would automatically be retained.
Next generation BI analytics is about governed self-service BI. Now a days if you imagine a visual in you head and you want to create it you need to go to expert for that. These are some of the visuals we created in PowerBI. The next generation is that these would be created by you.
http://www.nytimes.com/interactive/2012/02/13/us/politics/2013-budget-proposal-graphic.html?_r=0