🔍
Solution Accelerator Demo — ESG Analytics - YouTube
Channel: unknown
[1]
welcome to the databricks solution
[3]
accelerator demos
[4]
your fast preview on how to apply these
[6]
pre-built notebooks based on best
[8]
practices to solve common business
[9]
problems
[11]
today we will be looking at the solution
[12]
accelerator for environmental
[14]
social and governance better known as
[16]
esg analytics
[18]
esg is a data and ai challenge the world
[20]
of esg is highly unstructured by nature
[23]
if we look at the 40 most commonly
[24]
disclosed policies only 10 are metrics
[27]
and only 10 are hard numbers
[28]
the rest are purely policies initiatives
[31]
and mainly texts coming from different
[32]
systems
[34]
so how do we apply ai to quantify the
[36]
unquantifiable
[37]
and compare organizations in a much more
[39]
data-driven way than a subjective score
[42]
in a way that feeds into many different
[44]
use cases down the line from responsible
[46]
investing to supply chain resilience
[48]
carbon footprint reduction and even
[49]
reputational risk the lake house enables
[52]
processing of unstructured data such as
[53]
pdfs combined with ai capabilities
[56]
in the first series of esg notebooks we
[58]
show you how to extract those esg
[60]
initiatives
[61]
from unstructured pdf documents using ai
[64]
a solution that bridges the gap between
[65]
what a company says versus what a
[67]
company does
[68]
using alternative data we apply this
[71]
solution on news analytics and social
[73]
media to bridge the gap
[74]
and create a holistic view of an esg
[76]
practice that you can trust and act upon
[79]
this first notebook shows practitioners
[80]
how to programmatically extract pdf
[82]
documents
[83]
extract each and every single sentence
[85]
every single initiative
[86]
and apply natural language processing
[88]
and ai to understand what those
[90]
statements are all about
[91]
learning a vocabulary that is esg
[93]
specific with themes such as diversity
[95]
and inclusion code of conduct supporting
[97]
communities and green energy being able
[100]
to machine learn these themes and move
[102]
away from lengthy and verbose commonly
[104]
discussed policies to machine learned
[105]
initiatives that help us to summarize
[107]
complex pdf documents into 24 machine
[110]
learning policies that we can compare
[112]
we can also compare organizations side
[114]
by side based on how much they disclose
[116]
in each of those categories
[117]
if we take an example of a specific
[119]
industry or across your investments or
[121]
for your different competitors or your
[122]
different suppliers
[124]
we can look at how much company a says
[125]
they're valuing employees compared to
[127]
company b
[128]
how much more company c invested in
[130]
renewable energy compared to company d
[132]
and so on this provides an analytical
[134]
framework to help better understand
[136]
these organizations and how they differ
[137]
in their esg activities
[139]
by applying this model in news analytics
[141]
we show you how to bring 100 million
[142]
news articles in real time to understand
[145]
not only what a company is saying
[146]
but also what a company does
[148]
additionally we also find what the
[149]
reception was from the media and social
[151]
networks
[152]
with news and media analysis as a proxy
[154]
we show you how to extract
[156]
each and every single article across the
[157]
e the s and the g
[159]
and bring that real-time view of esg
[161]
instead of waiting for year-long csr
[163]
disclosures
[164]
this can be applied for every single
[165]
business large or medium companies
[167]
financial services or health care and
[169]
understanding the influence one
[171]
business's esg negative or positive
[173]
practices has on another in an
[174]
interconnected global market
[176]
moreover how can you act on those
[178]
different insights using a market risk
[180]
calculation looking at this from a
[181]
reputational risk or supply chain
[183]
resilience perspective
[184]
in this example we show you how to
[186]
create a simple synthetic portfolio
[188]
and tie those esg insights within the
[190]
market risk framework
[191]
it is no longer about how good a company
[193]
looks but also how performant this
[195]
company may be
[196]
and confirm what we know from research
[197]
and literature that shows that companies
[199]
with high esg practices tend to have
[201]
lower market volatility
[203]
we show in the context of value at risk
[205]
that an all synthetic portfolio is two
[207]
times more volatile due to bad esg
[209]
practices
[210]
finally we are able to bridge the gap
[212]
between the data science and engineering
[214]
world into a bi and ai dashboard
[216]
combining all those insights into one
[218]
platform to understand how much a
[220]
company says about e
[221]
s and g versus how much a company does
[223]
across each of those 24 machine learned
[225]
policies
[226]
informs us in real time about news
[228]
events that may positively or negatively
[230]
affect the esg of every single company
[232]
and in turn the impact it may have on
[234]
market performance
[236]
ready to get started on esg analytics
[238]
click on the link in the description
[239]
below to go right to our full write-up
[241]
for this solution and gain a deeper
[243]
understanding of how the data bricks
[244]
lake house uniquely solves for the
[246]
challenges associated with going from
[248]
batch to streaming
[249]
and bi to ai or visit the databrick
[252]
solution accelerator hub
[254]
to see all of our available accelerators
[255]
as well as keep up to date with new
[258]
launches
Most Recent Videos:
You can go back to the homepage right here: Homepage