# How do I identify strange data in access 2007

I have a collection of data where every now and then an incorrect setting somewhere in the data that was used previously to calculate the result caused the result to be wrong.

It is somewhat fortunate that when it goes wrong ... it goes real wrong .... so at a glance it is obvious something is wrong but the sheer amount of data makes it incredibly hard to find it in the first instance.

See the data I have provided.

With the 49 items given that should be similar with an average value of .228 it is obvious that the two numbers of 2600 are incorrect .... or at least different

I have a means of segregating the data into what should be like groups

Is there a way access can flag the two odd records?

Thanks
Paul ExampleData.xls
Asked:
###### Who is Participating?

Commented:
if you are expecting values less than one as correct values, you can use a query to pinpoint the wrong values

select product, totalcost
from table
where totalcost >1

0

Author Commented:
I am really trying to isolate data that is not the same.  It need not necessarily be any particular number as each "group" of data could have different values, minimums, maximums, averages, etc
0

Author Commented:
For example if I have 40 values between 1 and 2 .... I would not consider a value of 100 to be part of that set
0

Commented:
well, that will be difficult if you can not establish the correct parameters for pinpointing the incorrect values.
0

Author Commented:
Thinking back to my school days there was something called correlation that measured whether something was part of same group.

Too long ago
0

Commented:
You can continue study statistics, of course:)
But for simple analysis you can use stdev and avg.
In this example you can set quantity of stdev from avg for estimation of value.
DB26913114.zip
0

Commented:
>>Thinking back to my school days there was something called correlation that measured whether something
>>was part of same group.

That's not what correlation is.  Correlation quantifies the degree to which two random variables are related.

:)

Anyway, be very careful about using an approach that simply uses sample standard deviation and sample mean.  In your sample set, the sample standard deviation is 533, despite the fact that only two members of the set are >1.  Thus, if you set a rule such as "reject any items >2 SD from the mean", you could be asking for trouble.  (In a normal distribution, approximately 95% of the observations are expected to fall within 2 SD of the mean.)

Consider that if I add a single value of 300 to your set, the sample standard deviation stays high--about 528--and so this simple approach would accept 300 as valid, despite it being wholly unlike almost every other member of the set.

So, can you formulate a couple of rules that would identify a possible outlier?
0

Commented:
I imported the xls file into an Access table named tblCosts
Pick a number you want as the max value for the average under consideration. Then run this Query:

SELECT Avg(TotalCost) FROM tblCosts WHERE TotalCost < [EnterMaxValue];

0.3 yields 0.20738..
0.4 yields 0.21787..
0.5 yields 0.21816..

If all the data were together, you could enter both Min and Max values in the WHERE clause.  You get the idea.
0

Author Commented:
Thank you all very much for your suggestions and ideas.  I will muddle through.

I may be a little old for school

Cheers
Paul
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

## Already a member? Login.

All Courses

From novice to tech pro — start learning today.