Name: How Python transforms data analysis
Uploaded: 2024-02-09T10:31:05.885Z
Duration: 365.5
Description: Python language has been around for a while, but do you know how to leverage it for data analytics? Join us in the Streamly studio as Pierre Debois, CEO and Founder of Zimana Anlytics, makes Python make sense, whilst revealing how the language can supercharge your business insights team.

MIMS Python V3.mp4

MIMS Python V2.vtt

English

{{00:00:00}}
Hi, my name is Pierre DeBois.

{{00:00:01}}
I'm the founder and CEO
of Zimana Analytics.

{{00:00:04}}
I've been in business for 14 years
and I've worked with clients all

{{00:00:09}}
over helping them to understand
how to build data models.

{{00:00:31}}
Python is a open source objective
language, very popular for apps and

{{00:00:37}}
gaming, but in the data science world,
it's very popular for a number of

{{00:00:41}}
frameworks and applications to help
people understand their data and to

{{00:00:45}}
model their data in a statistical manner.

{{00:00:47}}
Python is used in a number of
frameworks ranging from TensorFlow

{{00:00:51}}
and PyTorch to even today when
we look at large language models.

{{00:00:55}}
For AI, many of them are
built with Python in mind.

{{00:01:04}}
Insights team may feel a little
bit overwhelmed with data.

{{00:01:07}}
So, one of the things that they can
use with Python initially is to do some

{{00:01:12}}
sort of cleanup and using that to get a
sense of how much data is really useful

{{00:01:16}}
for the model they intend to build.

{{00:01:18}}
There's a number of different frameworks
that allow users to identify, let's say,

{{00:01:23}}
non available data, maybe even outliers.

{{00:01:25}}
And to clean the data up such that it's
usable for a model, whether you're talking

{{00:01:30}}
about a machine learning model, maybe an
AI model through a large language model,

{{00:01:35}}
or even something just simpler in terms of
a data visualisation, trying to get a good

{{00:01:40}}
sense of what the data should look like.

{{00:01:41}}
Most

{{00:01:47}}
times, Python is put into application
for something statistical, and

{{00:01:51}}
here's what I mean by that.

{{00:01:52}}
If you're looking at data in, say,
Google Analytics, or maybe some,

{{00:01:57}}
some online analytics platform,
you would get a sense of trends and

{{00:02:02}}
sense of what things are repeating.

{{00:02:04}}
But to build an actual forecast
with a statistical model, you need,

{{00:02:10}}
you need Python and or frameworks
built around Python to do so.

{{00:02:14}}
The data that you have needs to have some
sort of statistical analysis to see if

{{00:02:19}}
you can actually create a correlation
between one said data versus the other

{{00:02:24}}
and a lot of frameworks  and modules have
been built for python can do that for you

{{00:02:34}}
Both have a little bit of a
of an interesting history.

{{00:02:37}}
I always always tell managers javascript
is almost like electricity in a car.

{{00:02:42}}
You need something to help put windows up
and down without turning the engine on.

{{00:02:45}}
And that's that's what
javascript was originally for

{{00:02:48}}
was for browsers helping to.

{{00:02:49}}
Have certain things move without having
to refresh the browser since then.

{{00:02:53}}
Developers have found
great uses for javascript.

{{00:02:56}}
But when it comes to data and
doing vectorisation, Python

{{00:03:00}}
is better suited for that.

{{00:03:01}}
There's modules that can help take that
data and create tensors, create all sorts

{{00:03:06}}
of data frames, all sorts of tables from
the data much more easily and with and

{{00:03:11}}
apply enough enough rigor to that data
to come up with models are  useful.

{{00:03:20}}
One of the difficulties with
Python, I think, comes more

{{00:03:24}}
with learning where to start.

{{00:03:27}}
Particularly for teams for corporate
teams, people hear Python all the

{{00:03:31}}
time, but they don't really hear about
the modules beyond NumPy and Panda.

{{00:03:36}}
Even starting with those two modules, you
can start to build up some understanding

{{00:03:40}}
of how to use Python for data models
in particular, and then learning

{{00:03:44}}
other frameworks that can go along
with it to accomplish certain tasks.

{{00:03:47}}
Doing that can help you craft
your projects not only much more

{{00:03:51}}
in a robust manner, but even
help your teams understand how

{{00:03:54}}
to use those tools properly.

{{00:03:56}}
The

{{00:04:01}}
data science professionals are
learning how Gen  AI is helping their

{{00:04:06}}
projects in terms of consolidating
advances that have been done from

{{00:04:11}}
Bard and with ChatGPT, particularly
for the advanced data analytics.

{{00:04:16}}
Has been really helpful in an open
up a number of options for people

{{00:04:21}}
to explore their data, but still up
to the data scientists to understand

{{00:04:26}}
when a model is hallucinating and to
put some domain knowledge to what's

{{00:04:30}}
being presented, so that way they
know how to make the adjustment and

{{00:04:33}}
even even asking the right prompts.

{{00:04:35}}
It's tied into the idea of domain
knowledge, knowing what to ask

{{00:04:40}}
and knowing what to look at and
having a sense of when to make the

{{00:04:43}}
adjustment and when to disregard.

{{00:04:51}}
I think when it comes to choosing
languages, there's always this sort of

{{00:04:54}}
tension between particularly between
Python and R, but there've been so much

{{00:04:58}}
development between the two of those.

{{00:05:00}}
I particularly like how Posit  has
positioned it's RStudio IDE for

{{00:05:06}}
example, to run Python and R.

{{00:05:09}}
I think it's, it's, it may be
more about preference these days.

{{00:05:12}}
There's always a counter argument for
using one language over the other.

{{00:05:16}}
So it's, it's probably more
about preference these days

{{00:05:19}}
more than anything else.

{{00:05:21}}
The

{{00:05:25}}
biggest advice I can give to any team
in terms of learning Python is to have

{{00:05:31}}
a really robust communication between
each other about skill sets and what

{{00:05:36}}
each member knows and not knows.

{{00:05:38}}
Python is a large language with a
lot of different frameworks, and

{{00:05:42}}
that means that a lot of different
domain knowledge can come into play.

{{00:05:46}}
Someone who is an expert in
visualisation will have a different

{{00:05:52}}
view of Python than someone who is
an expert on machine learning models.

{{00:05:57}}
Bringing those together
means communication.

I'm the founder and CEO
of Zimana Analytics.

I've been in business for 14 years
and I've worked with clients all

over helping them to understand
how to build data models.

Python is a open source objective
language, very popular for apps and

gaming, but in the data science world,
it's very popular for a number of

frameworks and applications to help
people understand their data and to

model their data in a statistical manner.

Python is used in a number of
frameworks ranging from TensorFlow

and PyTorch to even today when
we look at large language models.

For AI, many of them are
built with Python in mind.

Insights team may feel a little
bit overwhelmed with data.

So, one of the things that they can
use with Python initially is to do some

sort of cleanup and using that to get a
sense of how much data is really useful

There's a number of different frameworks
that allow users to identify, let's say,

And to clean the data up such that it's
usable for a model, whether you're talking

about a machine learning model, maybe an
AI model through a large language model,

or even something just simpler in terms of
a data visualisation, trying to get a good

times, Python is put into application
for something statistical, and

If you're looking at data in, say,
Google Analytics, or maybe some,

some online analytics platform,
you would get a sense of trends and

But to build an actual forecast
with a statistical model, you need,

you need Python and or frameworks
built around Python to do so.

The data that you have needs to have some
sort of statistical analysis to see if

you can actually create a correlation
between one said data versus the other

and a lot of frameworks  and modules have
been built for python can do that for you

Both have a little bit of a
of an interesting history.

I always always tell managers javascript
is almost like electricity in a car.

You need something to help put windows up
and down without turning the engine on.

And that's that's what
javascript was originally for

Have certain things move without having
to refresh the browser since then.

Developers have found
great uses for javascript.

But when it comes to data and
doing vectorisation, Python

There's modules that can help take that
data and create tensors, create all sorts

of data frames, all sorts of tables from
the data much more easily and with and

apply enough enough rigor to that data
to come up with models are  useful.

One of the difficulties with
Python, I think, comes more

Particularly for teams for corporate
teams, people hear Python all the

time, but they don't really hear about
the modules beyond NumPy and Panda.

Even starting with those two modules, you
can start to build up some understanding

of how to use Python for data models
in particular, and then learning

other frameworks that can go along
with it to accomplish certain tasks.

Doing that can help you craft
your projects not only much more

in a robust manner, but even
help your teams understand how

data science professionals are
learning how Gen  AI is helping their

projects in terms of consolidating
advances that have been done from

Bard and with ChatGPT, particularly
for the advanced data analytics.

Has been really helpful in an open
up a number of options for people

to explore their data, but still up
to the data scientists to understand

when a model is hallucinating and to
put some domain knowledge to what's

being presented, so that way they
know how to make the adjustment and

It's tied into the idea of domain
knowledge, knowing what to ask

and knowing what to look at and
having a sense of when to make the

I think when it comes to choosing
languages, there's always this sort of

tension between particularly between
Python and R, but there've been so much

I particularly like how Posit  has
positioned it's RStudio IDE for

I think it's, it's, it may be
more about preference these days.

There's always a counter argument for
using one language over the other.

So it's, it's probably more
about preference these days

biggest advice I can give to any team
in terms of learning Python is to have

a really robust communication between
each other about skill sets and what

Python is a large language with a
lot of different frameworks, and

that means that a lot of different
domain knowledge can come into play.

Someone who is an expert in
visualisation will have a different

view of Python than someone who is
an expert on machine learning models.

Bringing those together
means communication.

Make it Make Sense

Explainer

Python language has been around for a while, but do you know how to leverage it for data analytics? Join us in the Streamly studio as Pierre Debois, CEO and Founder of Zimana Anlytics, makes Python make sense, whilst revealing how the language can supercharge your business insights team. 

Machine Learning

Insights

Data analytics

data-driven decisions

Python

Large Language Model

thumbnail_Pierre DeBois_450844_1488223193.jpg

xsmall_Pierre DeBois_450844_1488223193.jpg

Pierre DeBois_450844_1488223193.jpg

Pierre DeBois

Founder / CEO

Debois teaches us what python is and how it gained it's popularity as a coding language.

The rise of Python

Most of the time Python is used for something analytical, Debois explains. But how else can the code language help your data insights team? 


Python use cases in data analytics

How can insights professionals decide which code language to use when there are so many good options? Debois looks at the advantages of Python over other languages.


Why Python and not JavaScript?

Debois shares his advice for teams wanting to integrate Python into their workflow.