I need some design help (time series analysis)

Hi,
the design looks ok, simple and straight forward, if you don=B4t want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--Hi,
Here is a fictious example that models my real life problem that I'm trying
to solve.
Lets say we have a bunch of baseball hitters and every day we give them an
hour to hit as many 'out of the park' home runs as they can hit. Now we want
to track home many they hit each day and watch whether their totals are
going up or down over time.
My fictious table design (as I see it) would be (with data)
PlayerId TotalHits Date
1 8 1/1/6
2 6 1/1/6
3 10 1/1/6
4 4 1/1/6
1 9 1/2/6
2 8 1/2/6
3 15 1/2/6
4 1 1/2/6
...
This may be poor design, please tell me if it is.
Anyway, from this design I could do a 'select playerid, sum(totalhits) from
xxx group by playerId' to see who has the most hits and look at each of the
top X hitters and see if they're in a incremental pattern or decreasing
pattern or just steady but high hits count type pattern.
Instead of that manual method I would like to figure a way to have the
system automatically tell me whose hit count is increasing over time and
whose are decreasing. and be able to break it down by week. month
quarter...
I'm guessing this might be analysis services in sql server, I've never used
them so I don't know. I run sql server 2000 btw.
If you have articles or references please point me in the right direction.
thanks|||Hi,
the design looks ok, simple and straight forward, if you don=B4t want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--|||"Jens" <Jens@.sqlserver2005.de> wrote in message
news:1151300420.928006.36970@.m73g2000cwd.googlegroups.com...
Hi,
the design looks ok, simple and straight forward, if you dont want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--
Ultimately I would like to see which players hit counts are on the rise and
which are on the decline. And yes I would like to be able to adjust the time
frames to analyze. So to start out I would like to see the query which would
show me the player id's whose hit counts are on the rise over the entire
time period stored in the database.
I can sort of see how the query could group by week (or month) but I don't
see how it can show me counts which are on the rise.
Thanks for your help.|||"Jens" <Jens@.sqlserver2005.de> wrote in message
news:1151300420.928006.36970@.m73g2000cwd.googlegroups.com...
Hi,
the design looks ok, simple and straight forward, if you dont want to
store more information here, you should be fine with that. Sure AS
could help you in some cases, but this should be just a simple query
here. Lets try to find a suggestion which will fit your needs. What do
you want to see at the end of the day, a list with playids only ? The
sums of their hits ? The difference between nother time period ? Which
time period do you want to investigate ? Try to give some sample
resultset that we can suggest something as an query for you.
HTH, Jens Suessmeyer
http://www.sqlserver2005.de
--
Ultimately I would like to see which players hit counts are on the rise and
which are on the decline. And yes I would like to be able to adjust the time
frames to analyze. So to start out I would like to see the query which would
show me the player id's whose hit counts are on the rise over the entire
time period stored in the database.
I can sort of see how the query could group by week (or month) but I don't
see how it can show me counts which are on the rise.
Thanks for your help.