That will will first cycle finalists in

the second final push to muster
enough vertical orders providing.

Also a voice for
the requests by the group.

Absolutely.

Thank you Joel.

Not the circumstances for the coup but

like the project was quite different from
that for there have been a constitutional.

They're good but
like improv it was forecast.

Still to test your view that I was trying
to stir forecasting as hard especially

about the future we are here in this group
and the forecast not only for the future.

They even had a hard time to forecast for
the Fast and the and

the presentation of the exchange
that explained why it was hard and

this for this project even to figure
out what was going on in the US and.

Some of the.

Obstacles to overcome.

One of the worst was obstacles
that lead to date of

presentational explain a little bit.

Now having problems with a lot of
nothing unusual in the senior design.

In fact I think most of your side projects
have to do with the lead on most of them.

I struggle with.

But there's a group of them that
really outstanding job in overcoming

the obstacles and
still produce interesting results and

I've produced interesting forecasting
results that also learned some valuable

lessons regarding their improvement in
the business process of this country and

they will tell us a little
bit of an aura of feeling.

I wanted to use to you.

They're kind of like the team to.

Good evening wear the COOPER Why
do you do you mind my work and

me Robbins and we are joined
by our other group members or

not you know you shuffle and
our daughter and son

in to nominate you for
my experience of twenty eight percent for

one of the largest cities with
the recent downturn in the economy.

The fact that Cooper has no
statistical models that take into

account economic factors must
really be extremely problematic for

there is in order to address this
problem are considered several

factors that may help
some particular sales.

In order to determine which of these
factors we perform Russian analysis and

the results of this analysis or
some models forecasting models in

the software package which is our
main We estimate the top savings

to be a little over one hundred
thousand dollars a year.

Cooper Lighting is headquartered here and

they're one of the largest manufacturers
of lighting products in the United States.

However there have been several different
markets for residential commercial and

also safe universe.

Such as exit lights and
outdoor lighting such as street lights.

They also sell for
five major sales channels channel where

they sell the products to distributors
who didn't sell goods to contractors and

also to the retail channel where they sell
for stores like Home Depot or Lowes and

these to mention are the two largest
sales channels in the project.

Some are next.

Coover it's not.

Contrary changes in the economy.

And therefore in office forecasts.

And they are based on intuition.

The management teams are also suffering
to have private business and so

there is no such a lot of systems.

This is what some of the problems
problems associated budgeting and

also increasing inventories.

So first I am talking about budget.

I first like to draw your attention to
the graph axis you can see yours and

the vertical axis you see the sales and
millions of dollars.

As you can see the lens original budget is
the actual sales that same time period.

Consistently over budget.

And this has led to increasing
errors in their budget.

They're incorrectly
allocating their resources.

This is really tying up a lot of aspects
of their business in order to account for

this it seems are having to perform
several I just missed throughout the year

and this is taking a lot
of unnecessary time.

Another problem is related
to increasing even Tori's.

I also like to try as much
to describe where the axis

again represents the time in years and
the furball access represents

the dollar value of goods being
held in the Tories the red and

the blue lines correspond to artist
friends the halo brand and model of spray.

And these two brands account for

approximately fifty percent of
the company sales of any given year and

this is why these two variants
were the focus of our project.

So inventories are increased
are consistently rising and

then you'll notice that there
are drop offs these drops are due to

write downs that management
have to perform as new

technologies are emerging that are more
energy efficient and is becoming obsolete.

There are forming right now five or
strappy.

Completely by reducing
the price significantly so

that they can sell all these kits.

And sell these two problems and

much as I mentioned are really just a lot
of companies resources and time and money.

As Melissa just outlined has a problem
with their forecasting to a variety

of problems in order to address and
we develop separate models for products.

And what is really a higher
construction of the products and

it's the lowest level that we can
get down to So for example these.

Might be something with the same
lights with a white background and

black background.

And this is the level
that we get older models.

The factors we consider more time.

Right.

And I think there's over
a hundred fifty factors.

And the economy and so we go into
the time factors that we considered.

And when we first started our project
believe that there would be a great

business because it was
a construction related business but

it turned out that there were
like significant in the models.

Patterns having to do with the corners
that will draw your attention to here.

You see this graph on the bottom.

We have separate quarters also
identified by the winds and

then you can see that in this
particular product family.

How the peaks correspond to the lines
which is the change in quarter families.

And then the next one.

If you look at another family
in a completely different

brand you can see the same exact
change with the peaks and the drops.

In order to capture this trend that
we saw in the business quarter

and a quarter indicator models.

And it just would not be able to meet

their quarter number at the end
when they realize they were either

drop their prices to meet that number but
then the distributors had already stopped

up they would have ordered the next
month and then it would drop.

So now going into price the price as
was mentioned earlier we had a lot of

problems with our data and this is
mainly were all of our lives in price.

So first like to graph pricing
distribution and what this is just

the number of transactions on the access
and the price of the transaction or.

First you can see how there's a lot of
products being sold at zero dollars

around zero dollars and
this has to do with bundling and.

It's just a bundle of products for
a certain price and

then when it's time to go back to the
computer to log the price of your product

they might already be finished the list.

So all the products after that they
want to sign a penny or five cents and

this is really inaccurate because that
wasn't accurate price of the product we

were able to just by getting those prices
from our average price calculator.

That you can see that wasn't as easy
to deal with and this was a tough.

About ten percent of the transactions
occurred above twenty dollars

all the way to two hundred forty when
there's a ton of prices over there at five

dollars around the ME THIS HAS TO DO WITH
first inconsistent.

They don't work directly for some within
one product would sell something.

For ten dollars and then the other
that same product for five dollars.

This was causing
the distributions to be crazy.

And then the second problem product.

I mentioned before how we felt that all
of the similar maybe just a different.

Or just otherwise family something.

For three dollars so
obviously these probably believed.

However our president wasn't terrible and

point out something that proved
to be very helpful for models.

Where most of the data
lies between the first

correspondents and
one channel the retail sales channel.

And the second or commercial and
industrial channels and

the very differently and actually very
beneficial for us to make separate models.

We have ten families making
twenty models and one for each.

And in the last group which was by far
the largest group or economic factors.

Of a hundred economic
factors a lot of things

considerably but we have the X.

and the economic factors
that we considered.

Brands something that was
very significant model's

brand so construction was very
significant also factors such as the S.

and P. five hundred of the stock market
and the consumer price index which is like

an inflation indicator and then once we
have those factors we consider separate

times to find which ones work for
each product family.

Month by month basis for
eighteen months in

advance and went back to

basics and figure out from there.

How different is

this offer different times.

As you can see here and

that time of sales dollars The horizontal

line represents five thousand nine

hundred percent

of the time

along the green line

just before the time period for

the same time period as you can see red.

Under ninety percent this

company they would have thirty
percent time they spent

their time spent about

one hundred five thousand dollars for
a private

private their recommendations for
our project

the first thirty years.

They're not there.

They're not there and also time

in better understand
there are probably other

better process

models se with these.

And also to the public.

So in summary for.

I am.