From: dave@davebolt.co.uk
To: 'Govardhan .G' <govardhansrr@gmail.com>
To: pgsql-admin@lists.postgresql.org
Subject: RE: Forcasting in Postgresql
Date: Thu, 2 Mar 2023 17:47:19 -0000
Message-ID: <020b01d94d2f$0933ed60$1b9bc820$@co.uk> (raw)
In-Reply-To: <CACpj7HWp-Lr-kVn1G0AqzG3xy_e5W5g5SyGX2hn-=9_eKGzdyA@mail.gmail.com>
References: <CACpj7HWp-Lr-kVn1G0AqzG3xy_e5W5g5SyGX2hn-=9_eKGzdyA@mail.gmail.com>
The short, unhelpful, answer is Yes.
Time to take a step back from the SQL and ask yourself:
Can I write an algorithm to forecast some value dependant on values I already have.
Very simple, somewhat vague, example.
I have been selling Snow Tyres for some time so I have values for sales along with dates.
What are the average sales for each month?
If I have more than one year’s data, what are the averages per month over the last X years?
Is a running average over the last few weeks/months/year any use?
Some made up figures.
Given that my average sales (in the Northern Hemisphere), were 20 in November, 18 in December, 2 in January, 3 in February, I might expect significant sales in November and December, but a drop in sales in January and February.
If last year the actual figures for November to February were 2,30,0,3, and the previous year were 38,6,4,3 then there is still a drop in sales in January and February, and still high sales in November and December, it’s just a little more complicated.
(If I were selling Snow Tyres, I would possibly look at these figures and get some stock in for November so that I would have stock on hand for the rest of the year, but now we are getting on to questions relating to order lead times).
I noticed that what I had intended to be a simple example quickly started to become more complicated.
I needed a clear definition of what I had, and what I wanted, so I tried to draw on my experience using a Stock Control System in the ‘80s and make up some easy data. I know that real systems that attempt to predict ordering quantities can soon get deep into Statistics, which is not one of my strengths. For this particular example I would be looking at search results for ‘stock management algorithm’ as a starting point.
Hope this helps a little.
Regards
Dave Bolt
From: Govardhan .G [mailto:govardhansrr@gmail.com]
Sent: 01 March 2023 07:48
To: pgsql-admin@lists.postgresql.org
Subject: Forcasting in Postgresql
Hi Folks,
Is it possible to write a query for forecasting in postgresql. I have been trying to search but have no clue. please advise
Thanks
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-admin@postgresql.org
Cc: dave@davebolt.co.uk, govardhansrr@gmail.com, pgsql-admin@lists.postgresql.org
Subject: RE: Forcasting in Postgresql
In-Reply-To: <020b01d94d2f$0933ed60$1b9bc820$@co.uk>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox