WEBVTT

00:00:00.760 --> 00:00:03.640
This is what my last week looked like.

00:00:04.680 --> 00:00:06.376
What I did,

00:00:06.400 --> 00:00:08.216
who I was with,

00:00:08.240 --> 00:00:11.360
the main sensations I had
for every waking hour ...

00:00:12.080 --> 00:00:14.576
If the feeling came as I thought of my dad

00:00:14.600 --> 00:00:16.376
who recently passed away,

00:00:16.400 --> 00:00:20.456
or if I could have just definitely
avoided the worries and anxieties.

00:00:20.480 --> 00:00:22.416
And if you think I'm a little obsessive,

00:00:22.440 --> 00:00:24.296
you're probably right.

00:00:24.320 --> 00:00:26.576
But clearly, from this visualization,

00:00:26.600 --> 00:00:29.896
you can learn much more about me
than from this other one,

00:00:29.920 --> 00:00:32.896
which are images you're
probably more familiar with

00:00:32.920 --> 00:00:35.736
and which you possibly even have
on your phone right now.

00:00:35.760 --> 00:00:37.976
Bar charts for the steps you walked,

00:00:38.000 --> 00:00:40.376
pie charts for the quality
of your sleep --

00:00:40.400 --> 00:00:42.120
the path of your morning runs.

NOTE Paragraph

00:00:43.280 --> 00:00:45.576
In my day job, I work with data.

00:00:45.600 --> 00:00:48.096
I run a data visualization design company,

00:00:48.120 --> 00:00:51.456
and we design and develop ways
to make information accessible

00:00:51.480 --> 00:00:53.576
through visual representations.

00:00:53.600 --> 00:00:56.616
What my job has taught me over the years

00:00:56.640 --> 00:01:00.856
is that to really understand data
and their true potential,

00:01:00.880 --> 00:01:03.976
sometimes we actually
have to forget about them

00:01:04.000 --> 00:01:05.760
and see through them instead.

00:01:06.440 --> 00:01:10.016
Because data are always
just a tool we use to represent reality.

00:01:10.040 --> 00:01:12.896
They're always used
as a placeholder for something else,

00:01:12.920 --> 00:01:15.096
but they are never the real thing.

NOTE Paragraph

00:01:15.120 --> 00:01:17.056
But let me step back for a moment

00:01:17.080 --> 00:01:19.560
to when I first understood
this personally.

00:01:20.280 --> 00:01:23.616
In 1994, I was 13 years old.

00:01:23.640 --> 00:01:25.576
I was a teenager in Italy.

00:01:25.600 --> 00:01:28.216
I was too young
to be interested in politics,

00:01:28.240 --> 00:01:30.736
but I knew that a businessman,
Silvio Berlusconi,

00:01:30.760 --> 00:01:33.320
was running for president
for the moderate right.

00:01:34.120 --> 00:01:36.136
We lived in a very liberal town,

00:01:36.160 --> 00:01:39.456
and my father was a politician
for the Democratic Party.

00:01:39.480 --> 00:01:43.816
And I remember that no one thought
that Berlusconi could get elected --

00:01:43.840 --> 00:01:45.600
that was totally not an option.

00:01:46.560 --> 00:01:47.776
But it happened.

00:01:47.800 --> 00:01:50.336
And I remember the feeling very vividly.

00:01:50.360 --> 00:01:52.216
It was a complete surprise,

00:01:52.240 --> 00:01:57.280
as my dad promised that in my town
he knew nobody who voted for him.

NOTE Paragraph

00:01:58.720 --> 00:02:00.216
This was the first time

00:02:00.240 --> 00:02:04.800
when the data I had gave me
a completely distorted image of reality.

00:02:05.280 --> 00:02:08.576
My data sample was actually
pretty limited and skewed,

00:02:08.600 --> 00:02:12.336
so probably it was because of that,
I thought, I lived in a bubble,

00:02:12.360 --> 00:02:14.960
and I didn't have enough chances
to see outside of it.

NOTE Paragraph

00:02:16.080 --> 00:02:19.856
Now, fast-forward to November 8, 2016

00:02:19.880 --> 00:02:21.080
in the United States.

00:02:21.960 --> 00:02:23.256
The internet polls,

00:02:23.280 --> 00:02:24.656
statistical models,

00:02:24.680 --> 00:02:29.496
all the pundits agreeing on a possible
outcome for the presidential election.

00:02:29.520 --> 00:02:32.136
It looked like we had
enough information this time,

00:02:32.160 --> 00:02:36.256
and many more chances to see outside
the closed circle we lived in --

00:02:36.280 --> 00:02:37.600
but we clearly didn't.

00:02:38.040 --> 00:02:40.136
The feeling felt very familiar.

00:02:40.160 --> 00:02:41.640
I had been there before.

00:02:42.360 --> 00:02:45.216
I think it's fair to say
the data failed us this time --

00:02:45.240 --> 00:02:47.096
and pretty spectacularly.

00:02:47.120 --> 00:02:48.816
We believed in data,

00:02:48.840 --> 00:02:50.256
but what happened,

00:02:50.280 --> 00:02:52.976
even with the most respected newspaper,

00:02:53.000 --> 00:02:57.696
is that the obsession to reduce everything
to two simple percentage numbers

00:02:57.720 --> 00:02:59.696
to make a powerful headline

00:02:59.720 --> 00:03:01.776
made us focus on these two digits

00:03:01.800 --> 00:03:03.000
and them alone.

00:03:03.560 --> 00:03:05.616
In an effort to simplify the message

00:03:05.640 --> 00:03:09.056
and draw a beautiful,
inevitable red and blue map,

00:03:09.080 --> 00:03:10.960
we lost the point completely.

00:03:11.440 --> 00:03:13.576
We somehow forgot
that there were stories --

00:03:13.600 --> 00:03:15.960
stories of human beings
behind these numbers.

NOTE Paragraph

00:03:17.240 --> 00:03:18.816
In a different context,

00:03:18.840 --> 00:03:20.496
but to a very similar point,

00:03:20.520 --> 00:03:24.416
a peculiar challenge was presented
to my team by this woman.

00:03:24.440 --> 00:03:26.816
She came to us with a lot of data,

00:03:26.840 --> 00:03:31.256
but ultimately she wanted to tell
one of the most humane stories possible.

00:03:31.280 --> 00:03:32.976
She's Samantha Cristoforetti.

00:03:33.000 --> 00:03:35.576
She has been the first
Italian woman astronaut,

00:03:35.600 --> 00:03:38.096
and she contacted us before being launched

00:03:38.120 --> 00:03:42.016
on a six-month-long expedition
to the International Space Station.

00:03:42.040 --> 00:03:44.256
She told us, "I'm going to space,

00:03:44.280 --> 00:03:47.376
and I want to do something meaningful
with the data of my mission

00:03:47.400 --> 00:03:48.640
to reach out to people."

00:03:49.600 --> 00:03:52.136
A mission to the
International Space Station

00:03:52.160 --> 00:03:54.256
comes with terabytes of data

00:03:54.280 --> 00:03:56.656
about anything you can possibly imagine --

00:03:56.680 --> 00:03:58.176
the orbits around Earth,

00:03:58.200 --> 00:04:00.296
the speed and position of the ISS

00:04:00.320 --> 00:04:04.000
and all of the other thousands
of live streams from its sensors.

00:04:04.840 --> 00:04:07.736
We had all of the hard data
we could think of --

00:04:07.760 --> 00:04:10.176
just like the pundits
before the election --

00:04:10.200 --> 00:04:13.176
but what is the point
of all these numbers?

00:04:13.200 --> 00:04:15.936
People are not interested
in data for the sake of it,

00:04:15.960 --> 00:04:17.815
because numbers are never the point.

00:04:17.839 --> 00:04:19.800
They're always the means to an end.

00:04:20.839 --> 00:04:22.616
The story we needed to tell

00:04:22.640 --> 00:04:25.136
is that there is a human being
in a teeny box

00:04:25.160 --> 00:04:27.416
flying in space above your head,

00:04:27.440 --> 00:04:31.536
and that you can actually see her
with your naked eye on a clear night.

00:04:31.560 --> 00:04:34.656
So we decided to use data
to create a connection

00:04:34.680 --> 00:04:38.736
between Samantha and all of the people
looking at her from below.

00:04:38.760 --> 00:04:41.936
We designed and developed
what we called "Friends in Space,"

00:04:41.960 --> 00:04:46.616
a web application that simply
lets you say "hello" to Samantha

00:04:46.640 --> 00:04:47.896
from where you are,

00:04:47.920 --> 00:04:51.456
and "hello" to all the people
who are online at the same time

00:04:51.480 --> 00:04:53.000
from all over the world.

00:04:53.640 --> 00:04:57.096
And all of these "hellos"
left visible marks on the map

00:04:57.120 --> 00:04:59.136
as Samantha was flying by

00:04:59.160 --> 00:05:02.536
and as she was actually
waving back every day at us

00:05:02.560 --> 00:05:04.240
using Twitter from the ISS.

NOTE Paragraph

00:05:04.880 --> 00:05:09.856
This made people see the mission's data
from a very different perspective.

00:05:09.880 --> 00:05:14.576
It all suddenly became much more
about our human nature and our curiosity,

00:05:14.600 --> 00:05:16.256
rather than technology.

00:05:16.280 --> 00:05:18.616
So data powered the experience,

00:05:18.640 --> 00:05:21.040
but stories of human beings
were the drive.

00:05:22.840 --> 00:05:26.176
The very positive response
of its thousands of users

00:05:26.200 --> 00:05:28.136
taught me a very important lesson --

00:05:28.160 --> 00:05:31.016
that working with data
means designing ways

00:05:31.040 --> 00:05:33.776
to transform the abstract
and the uncountable

00:05:33.800 --> 00:05:37.816
into something that can be seen,
felt and directly reconnected

00:05:37.840 --> 00:05:40.136
to our lives and to our behaviors,

00:05:40.160 --> 00:05:42.016
something that is hard to achieve

00:05:42.040 --> 00:05:45.936
if we let the obsession for the numbers
and the technology around them

00:05:45.960 --> 00:05:47.240
lead us in the process.

00:05:48.600 --> 00:05:53.496
But we can do even more to connect data
to the stories they represent.

00:05:53.520 --> 00:05:56.176
We can remove technology completely.

NOTE Paragraph

00:05:56.200 --> 00:05:58.456
A few years ago, I met this other woman,

00:05:58.480 --> 00:05:59.856
Stefanie Posavec --

00:05:59.880 --> 00:06:05.696
a London-based designer who shares with me
the passion and obsession about data.

00:06:05.720 --> 00:06:07.056
We didn't know each other,

00:06:07.080 --> 00:06:10.336
but we decided to run
a very radical experiment,

00:06:10.360 --> 00:06:12.896
starting a communication using only data,

00:06:12.920 --> 00:06:14.256
no other language,

00:06:14.280 --> 00:06:18.896
and we opted for using no technology
whatsoever to share our data.

00:06:18.920 --> 00:06:21.816
In fact, our only means of communication

00:06:21.840 --> 00:06:24.696
would be through
the old-fashioned post office.

00:06:24.720 --> 00:06:27.176
For "Dear Data," every week for one year,

00:06:27.200 --> 00:06:30.656
we used our personal data
to get to know each other --

00:06:30.680 --> 00:06:34.336
personal data around weekly
shared mundane topics,

00:06:34.360 --> 00:06:35.576
from our feelings

00:06:35.600 --> 00:06:37.456
to the interactions with our partners,

00:06:37.480 --> 00:06:40.640
from the compliments we received
to the sounds of our surroundings.

00:06:41.480 --> 00:06:45.016
Personal information
that we would then manually hand draw

00:06:45.040 --> 00:06:47.536
on a postcard-size sheet of paper

00:06:47.560 --> 00:06:50.496
that we would every week
send from London to New York,

00:06:50.520 --> 00:06:51.776
where I live,

00:06:51.800 --> 00:06:54.000
and from New York to London,
where she lives.

00:06:54.480 --> 00:06:58.176
The front of the postcard
is the data drawing,

00:06:58.200 --> 00:06:59.496
and the back of the card

00:06:59.520 --> 00:07:01.949
contains the address
of the other person, of course,

00:07:01.973 --> 00:07:04.613
and the legend for how
to interpret our drawing.

00:07:05.640 --> 00:07:07.656
The very first week into the project,

00:07:07.680 --> 00:07:10.736
we actually chose
a pretty cold and impersonal topic.

00:07:10.760 --> 00:07:13.960
How many times do we
check the time in a week?

00:07:14.720 --> 00:07:16.656
So here is the front of my card,

00:07:16.680 --> 00:07:18.656
and you can see that every little symbol

00:07:18.680 --> 00:07:22.096
represents all of the times
that I checked the time,

00:07:22.120 --> 00:07:25.496
positioned for days
and different hours chronologically --

00:07:25.520 --> 00:07:27.560
nothing really complicated here.

00:07:28.200 --> 00:07:29.776
But then you see in the legend

00:07:29.800 --> 00:07:33.256
how I added anecdotal details
about these moments.

00:07:33.280 --> 00:07:37.856
In fact, the different types of symbols
indicate why I was checking the time --

00:07:37.880 --> 00:07:39.096
what was I doing?

00:07:39.120 --> 00:07:40.816
Was I bored? Was I hungry?

00:07:40.840 --> 00:07:42.056
Was I late?

00:07:42.080 --> 00:07:45.296
Did I check it on purpose
or just casually glance at the clock?

00:07:45.320 --> 00:07:47.576
And this is the key part --

00:07:47.600 --> 00:07:51.296
representing the details
of my days and my personality

00:07:51.320 --> 00:07:53.256
through my data collection.

00:07:53.280 --> 00:07:57.976
Using data as a lens or a filter
to discover and reveal, for example,

00:07:58.000 --> 00:08:00.176
my never-ending anxiety for being late,

00:08:00.200 --> 00:08:02.400
even though I'm absolutely always on time.

NOTE Paragraph

00:08:04.200 --> 00:08:08.296
Stefanie and I spent one year
collecting our data manually

00:08:08.320 --> 00:08:12.816
to force us to focus on the nuances
that computers cannot gather --

00:08:12.840 --> 00:08:14.376
or at least not yet --

00:08:14.400 --> 00:08:17.896
using data also to explore our minds
and the words we use,

00:08:17.920 --> 00:08:19.856
and not only our activities.

00:08:19.880 --> 00:08:21.296
Like at week number three,

00:08:21.320 --> 00:08:25.136
where we tracked the "thank yous"
we said and were received,

00:08:25.160 --> 00:08:29.816
and when I realized that I thank
mostly people that I don't know.

00:08:29.840 --> 00:08:34.176
Apparently I'm a compulsive thanker
to waitresses and waiters,

00:08:34.200 --> 00:08:37.360
but I definitely don't thank enough
the people who are close to me.

NOTE Paragraph

00:08:39.000 --> 00:08:40.256
Over one year,

00:08:40.280 --> 00:08:44.776
the process of actively noticing
and counting these types of actions

00:08:44.800 --> 00:08:46.096
became a ritual.

00:08:46.120 --> 00:08:48.176
It actually changed ourselves.

00:08:48.200 --> 00:08:50.896
We became much more
in tune with ourselves,

00:08:50.920 --> 00:08:54.040
much more aware of our behaviors
and our surroundings.

00:08:54.680 --> 00:08:57.656
Over one year, Stefanie and I
connected at a very deep level

00:08:57.680 --> 00:08:59.696
through our shared data diary,

00:08:59.720 --> 00:09:04.016
but we could do this only because
we put ourselves in these numbers,

00:09:04.040 --> 00:09:08.016
adding the contexts
of our very personal stories to them.

00:09:08.040 --> 00:09:10.496
It was the only way
to make them truly meaningful

00:09:10.520 --> 00:09:12.720
and representative of ourselves.

NOTE Paragraph

00:09:14.480 --> 00:09:17.576
I am not asking you
to start drawing your personal data,

00:09:17.600 --> 00:09:20.456
or to find a pen pal across the ocean.

00:09:20.480 --> 00:09:23.056
But I'm asking you to consider data --

00:09:23.080 --> 00:09:24.536
all kind of data --

00:09:24.560 --> 00:09:26.336
as the beginning of the conversation

00:09:26.360 --> 00:09:27.560
and not the end.

00:09:28.080 --> 00:09:31.256
Because data alone
will never give us a solution.

00:09:31.280 --> 00:09:33.976
And this is why data failed us so badly --

00:09:34.000 --> 00:09:37.376
because we failed to include
the right amount of context

00:09:37.400 --> 00:09:38.856
to represent reality --

00:09:38.880 --> 00:09:42.080
a nuanced, complicated
and intricate reality.

00:09:42.960 --> 00:09:45.416
We kept looking at these two numbers,

00:09:45.440 --> 00:09:46.936
obsessing with them

00:09:46.960 --> 00:09:49.456
and pretending that our world
could be reduced

00:09:49.480 --> 00:09:51.816
to a couple digits and a horse race,

00:09:51.840 --> 00:09:53.096
while the real stories,

00:09:53.120 --> 00:09:54.576
the ones that really mattered,

00:09:54.600 --> 00:09:56.016
were somewhere else.

NOTE Paragraph

00:09:56.040 --> 00:10:00.456
What we missed looking at these stories
only through models and algorithms

00:10:00.480 --> 00:10:03.000
is what I call "data humanism."

00:10:03.560 --> 00:10:05.576
In the Renaissance humanism,

00:10:05.600 --> 00:10:07.216
European intellectuals

00:10:07.240 --> 00:10:12.160
placed the human nature instead of God
at the center of their view of the world.

00:10:12.800 --> 00:10:15.016
I believe something similar
needs to happen

00:10:15.040 --> 00:10:16.816
with the universe of data.

00:10:16.840 --> 00:10:19.816
Now data are apparently
treated like a God --

00:10:19.840 --> 00:10:23.120
keeper of infallible truth
for our present and our future.

NOTE Paragraph

00:10:23.840 --> 00:10:26.736
The experiences
that I shared with you today

00:10:26.760 --> 00:10:31.776
taught me that to make data faithfully
representative of our human nature

00:10:31.800 --> 00:10:35.216
and to make sure they will not
mislead us anymore,

00:10:35.240 --> 00:10:38.936
we need to start designing ways
to include empathy, imperfection

00:10:38.960 --> 00:10:40.536
and human qualities

00:10:40.560 --> 00:10:44.280
in how we collect, process,
analyze and display them.

00:10:45.280 --> 00:10:48.256
I do see a place where, ultimately,

00:10:48.280 --> 00:10:51.616
instead of using data
only to become more efficient,

00:10:51.640 --> 00:10:54.440
we will all use data
to become more humane.

NOTE Paragraph

00:10:54.880 --> 00:10:56.096
Thank you.

NOTE Paragraph

00:10:56.120 --> 00:11:00.561
(Applause)