Let’s take a small test. We’ll show a table of numbers on the screen, and ask 3 questions about those numbers. You have 30 seconds to answer these. You can just write down the answers or remember them – there’s no need to say the answer out aloud.
Your timer starts now.
What answers did you get?
How many numbers were above 100? How many were below 10? Which quadrant had the highest total?
[Typically, there will be a lot of variance in these answers] So there’s considerable variation in the answers you get.
Now, let’s do the same exercise again, but with some extremely simple highlighting. It’s the same questions. You have 30 seconds. This time, you can say the answer out aloud if you like.
Your time starts now.
You’ll now see a series of examples of the work Gramener has done with its clients.
All of these examples, every visual that you will see from now on (including the ones in PowerPoint) were directly created as output from the Gramener visualisation server.
We were also interested in applying these rich visualisations to sports. One question we had was, for example, “Who’s the fastest one day international player?”
The trouble with that is, depending on when you measure it and how you measure it, the results could be very different. For example, if we take strike rate as a metric, it turned out (when we did it) that it was a South African who had the highest strike rate – of 200%. He played one match, hit a four, and got out the next ball.
Clearly, that’s not what we’re looking for. We could, perhaps, take a minimum number of runs as a cut-off. But the question is, what should that be? 100? 1000? 5000? Where does one draw the line, and why is that the right one? If you don’t know the domain, answering this is difficult.
Like with the contract farming example before, we need a way of looking at performance combined with scale or importance.
We said, let’s take all of the players who’ve ever played one day internationals. Each box is one player. The size of the rectangle is proportional to the number of runs they’ve scored.
So you can see that Tendulkar has scored the most runs, followed by Ganguly, then Dravid, and so on. The size of the entire visual is representative of the total runs ever scored by India in one-day internationals.
Colour is based on the strike rate. The greener the rectangle, the faster the score. You can see that Sehwag has done a fairly good job. So has Yusuf Pathan, one of the smaller green boxes. But given that that box is just about 1/10th the size of Sehwag’s, you could say that Yusuf Pathan has a long way to go before he can be considered on par.
You’ll also find that many of the players who have a lower run rate – like Ravi Shastri, Dilip Vengsarkar, Sunil Gavaskar, Mohinder Amarnath, etc. – were playing in a different era, a time when a score of 200 was considered a rather good score. Today, 300 would be a respectable target.
It turns out that strike rate increases at around 3.5% every decade. If we adjust for that and re-plot the strike rates, it emerges that Kapil Dev’s adjusted strike rate is almost exactly the same as Sewhag’s, and between them, we have the two fastest players India has had.
Gramener is a data analytics and visualisation company.
We have the ability to process data at a small and a large scale. We analyse the data to find non-intuitive insights that lie hidden behind it and present it as a visual story that makes those insights obvious in real time.
Most discussions of decision-making assume,
that only senior executives make decisions or
that only senior executives’ decisions matter.
This is a dangerous mistake…
- Peter F. Drucker
It's clearly a budget! Has a lot of numbers in it!
- George W. Bush
Information is the oil of the 21st century,
and analytics is the combustion engine
Business analytics software grew 14% in
2011 and will hit $50.7 bn by 2016
… the #1 trend is applying information &
analytics to solve business problems
Increasing data being churned out
by systems in information highway
Social network data
Consumers embracing Web 2.0 and
the social media lifestyle
Portable devices generating data for
consumption by systems
Material science research has led to
significant increase in data density
Driven by massive investments in
Moore’s law has doubled the
processing power per $ every 1.5 yrs
Growth in available data, and the
potential for exploiting these,
have grown exponentially in the
last 10 years.
This changing data landscape
heralds a radical shift in business
decision-making approach, even
for mere survival in this new age.
Data growth to 7.9 ZB by
2015 posing a real ‘Data
Gartner’s BI Magic Quadrant Trends
• Emergence of data discovery/ visualization
• Increased willingness for new low-cost options
• Embedded low-cost purpose-built analytic apps
• Need for intuitive BI tools on mobile platforms
THE VISUAL INTELLIGENCE MARKET
Business Analytics software market is slated to
grow at 10% annually and hit $50.7 Bn by 2016
• Big Data market would be $16.9 Bn
• Data Discovery market would be $ 1.6 Bn
Today, only strategic decisions are made at a rate slower than the speed of business.
Tactical and operational decisions must increasingly be at a rate faster than humans
are capable of.
…challenge of mastering big data analytics is the hardest, because big data technologies
are not ready for enterprise demands, & skills to work with big data are scarce
Data discovery and visualization vendor initiatives, and the rapid adoption
from end users, have the potential to shake the BI market to its foundations.
“Visualization would be the platform on
which Big Data consumption would happen”
WHO USES DATA VISUALISATION?
New York Times
What determines a child’s marks?
Do girls score better than boys?
Does the choice of subject matter?
Does the medium of instruction matter?
Does community or religion matter?
Does their birthday matter?
Does the first letter of their name matter?
Based on the results of the 20 lakh
students taking the Class XII exams
at Tamil Nadu over the last 3 years,
it appears that the month you were
born in can make a difference of as
much as 120 marks out of 1,200.
score the lowest
The marks shoot
up for Aug borns
… and peaks for
120 marks out of
by month of birth
An identical pattern was observed in 2009 and 2010…
… and across districts, gender, subjects, and class X & XII.
“It’s simply that in Canada the eligibility
cutoff for age-class hockey is January 1. A
boy who turns ten on January 2, then,
could be playing alongside someone who
doesn’t turn ten until the end of the year—
and at that age, in preadolescence, a
twelve-month gap in age represents an
enormous difference in physical maturity.”
-- Malcolm Gladwell, Outliers
BIRTHDAYS IN THE US AND IN INDIA
I’ve always been curious… who
among India’s prolific one-day
run-getters had the best strike
What about the rest of the world?
INDIA ODI BATTING
Shift Evening Morning Night
Weekday Fri Mon Sat Sun Thu Tue Wed
Product category FAH N70 RPP TDS ZDH
Part shipment 20-40% 40-60% 60-80% <20% Full
This visualisation measures the recovery time (time from
arrival of the flight until delivery), and identifies which
factors most influence the recovery time.
Recovery times are neutral during the evening and morning shifts (mornings are slightly worse), night times are the best.
Recovery times are worst on Fridays, and best on Saturdays & Wednesdays.
Specifically, Friday mornings are particularly bad. So are Thursday mornings.
The FAH product category has the best recovery time, while ZDH is much worse.
However, RPP on Sundays is unusually slow.
Part shipped products tend to perform worse than full-shipments. Specifically the <20% and 40-60% part-shipments.
This is especially problematic for ZDH
This visualisation is part of a suite of analytical techniques we call “grouped
means” that allows us to measure the impact of every parameter (shifts,
weekdays, etc.) on any measure of interest – recovery time in this case, but this
could be extended to revenue, operational efficiency, or ability to cross-sell.
It allows automatically detection of
statistically significant flows and
highlights only relevant ones to users.
The system therefore analyses all
possible patterns, but users only see
the insights that matter.
THE SOCIAL TALE OF TWO CITIES: BANGALORE & SINGAPORE
Recruiting top quality developers is always a problem. We decided to use an
algorithmic approach and pulled out the social network of developers on
Github (a social network for open source code).
In this visualisation, each circle is a person. The size of the circle
represents the number of followers. Larger circles have more
followers (but not in proportion – it’s a log scale.)
The circle’s colour represents the city the
programmer’s live in. This visual is a slice showing the
tale of two cities: Bangalore and Singapore
Two people are connected if one
follows the other. This leads to a
clustering of people in the form of a
Here, you can see that Bangalore and
Singapore are reasonably well
connected cities. Bangalore has more
developers, but Singapore has more
popular ones (larger circles).
However, the interaction between
Bangalore and Singapore are few and
far between. But for a few people
across both cities, like:
Anand C, Consultant
Anand S, Gramener
Sau Sheong, HP Labs
Lim Chee Aung
A follows B (or)
B follows A
Most followed in
Most followed in
There are, of course, a number of smaller
independent circles – people who are not connected
to others in the same city. (They may be connected to
people in other cities.)
Apart from this, there are a few small networks of
connected people – often people within the same
company or start-up – who form a community of their
Can we visualize the results of
every single Lok Sabha election
We handle terabyte-size data via non-traditional analytics and visualise it in real-time.
Gramener transforms your data into concise dashboards
that make your business problem & solution visually obvious.
We help you find insights quickly, based on cognitive research,
and our visualisations guide you towards actionable decisions.
A data analytics and visualisation company
EDWARD TUFTE: CLASSICS ON VISUALIZATION
STEPHEN FEW: DASHBOARD DESIGN