I think we always have to think about the ethical side when we do data science we need to have a critical point of view, to take a step back. Artificial intelligence is a
science and like any science that respect itself, it may have
weak sides. Today we took the plane and we
are in Palo Alto, Los Angeles California, to talk about
artificial intelligence, big data data scientist …
It's almost that Miguel except that in fact we did not take a plane we are not in
California we are in Bordeaux in a building called the Data Factory
there are four companies here that study the data, it's a obscure concept for us so we met Paul, Jean Paul, who will explain us. At the end of the video you are going to understand what is a data scientist!
So that's the idea, he will explain us a little bit of all this:
what is big data, what is artificial intelligence and especially what is his job about and in which, in which areas
applies, in which industries can we work when we are a data scientist?
Will discover all this with him Come on, we're going back inside quickly? Hello my name is Jean Paul I'm 33 years old I am a data scientist. The job of
data scientist it's just helping the companies or researchers
scientists to make decisions to from the data they produce or
that they hold and so today "data science" is a word that
buzz on the internet and there are plenty of
words that make the buzz too and revolve around this notion:
so we will talk about artifial intelligence, we will talk about big data … the
big data contains techniques of perception, of storage and treatment
of colossal masses of data so imagine Facebook
who receives every day I do not know, 1 billion photos per day Facebook must learn to perceive these
data, to store them on servers farms and to treat them, to make
something. So today data scientists do not
do necessarily big data, there are very few companies that process their data today. And what is artificial intelligence?
Artificial intelligence brings together artificial calculation techniques that
do that mimic in somehow a cognitive ability or multiple cognitives abilities of the human. Why do we talk about artificial intelligence
around datas around the data? I do not know, we'll imagine Camille's Facebook profile on which she is going submit photos every day,
there are friends of hers who will tag her, who will tag her face so
Facebook is going to be pleased. The brain of Facebook will come to tell itself
"I have Camille faces who are identified as Camille, I'm going
to try to learn to recognize Camille on new photos ", and
this faculty of learning is a faculty
cognitive of the human. It is in this context that we speak of
artificial intelligence, including the techniques that we call "machine learning".
My job is to keep up up to day of the
new technologies coming out, bring me up to date with all the
new techniques that can help analyze data.
My job is to write a lot of computer code, especially in languages
R and python which are the two most used languages in data science programming today. I like pedagogy,
popularization, I like to transmit to others…
Yeah, I like it! The domains application that I prefer is the
areas of academic research, including research in medicine,
social sciences, agronomy … Ah, it's not over? My job requires
a lot of plasticity, curiosity. I can reach an immeasurable number of areas. What I like less in my job it may be the
domains, some areas of application that I appreciate a little less: in the world of consumption, especially targeted advertising … it sticks maybe a little bit less to my values. What does it look like a typical day?
It's a lot of time spent behind his screen to code
prediction or decision-making algorithms but also a lot of exchanges with several teams of the company:
the computer team for example but especially the decision-making team, the
marketing teams, the R & D teams, which have specific needs. Are there a lot of employment ? Yes, there are a lot employment options today. In fact in which activities sectors types can you work? Outside the (private) enterprise, in the world of
research we need a lot the data science. I like the agronomy world because that's where I come from We use
a lot data science in medicine, in epidemiology for example, in psychology … I think it's my favorite area, it's quite fascinating.
Ok it's still a bit abstract for me: What do you harvest
concretely as data and how do you harvest to be able to isolate a
gene and say that gene has such property ? Good I am going to give you a
quite concrete protocol: we are going take for example thirty
oysters that have some superior quality compared to about thirty
other oysters that have some quality or a phenotype if we have to use an intelligent word, a
lower quality. What we're going to harvest, it's information
about the genome of these two groups of oysters.
We will try to see what are the parts of the genome that change between
the two groups of oysters: we can assume that it is the genes that determine in
somehow the quality of oysters, which allows us to make oyster selections of better qualities. Does it seem less abstract? No no, yes! So finally we're going
talk about salary: I think the junior data scientists can
start or start on an average of 38K or 40K euros a year and there is no
limit for seniors it can go, it can go very far. With the excitement around today of the business of data science, there is a good
number of engineering schools and schools of business that have integrated the course of
data science in their program. In general data scientists come either from the
computer world, or the world of mathematics.
We have mathematicians who have acquired computer skills or
computer scientists who have acquired math skills and
statistics, who suddenly became data scientist. Otherwise we can become
data scientist self-taught: so I'll tell you about my background,
I started with a license in biology.
I did a master's degree in ecology, I was fascinated by the world of research
so I followed with on a thesis of doctorate in ecology. During my
thesis I had the opportunity to harvest a a lot of data on biodiversity, ecosystem functioning … and I had to learn to analyze them,
analyze this data to be able to draw conclusions of my research and
as nobody was available in my lab to teach me to
analyze I had to learn on my own,
searching on the internet, with tutorials, in
books … so it's done quite easily, there is enough documentation around that. I had the opportunity to intervene in
different research laboratories especially in France at the CNRS, INSERM, CEA, INRA to initiate researchers to data analysis.
I also had the opportunity with this job to create videos of
scientific popularization around data analysis.
A few months ago I decided to become
a freelance as an advisor and trainer in data analysis and in
statistics. Today we say in "data sciences", maybe
in a few years we will say something else ! We finished shooting the job of
data scientist. We hope you enjoyed it! It's a
job that has a lot of areas of application, which is going to be I think
more and more democratized, so if you ever have
people around you who are interested in this area of activity
do not hesitate to share our videos! As Jean Paul said, you can become a data scientist via any career. We leave you, we return
inside the building behind us, the data factory building! We are going to take a good coffee with Jean Paul and we say you until next week!