Yes thats correct, I am hoping that I will find software/addon that would think for me. However in programmed way. Even further - software/addon should have feature that works like ''human brains'' - particular filter's parameters shall be missing in the collection of BW addon. Example: for time perion (syntax: mm.yyyy) there might be only 02.2013, 03.2013, 05.2013 but if I define period from first month in 2013 to first month in 2014 the software/addon I am asking for should automatically ''know'' to include the other parameters such as 01.2013 (not yet added), 04.2013, 06.2013 and so on until, including (but for the values still available after refreshing file to most up to date version from people doing the upgrades of values), 01.2014. This is what I meant with ''human brains''.



It would save me/us a LOT of work so please don't just say nothing like this exist. There have to be some comparison tool that could work with BW customized isolated addon (but still located inside excel). Would save also months of time.

Give report out. Which report? Whatever I define as being interested for. But remember: I define what report I want and not what to compare.

I apologize but i don't understand neither of those two questions. Please clarify, perhaps on some example.

Most of your remaining, latest, reply is correct.

Since this conversation have been going on for a while, since I still don't have the name of the program and since Im aware the urgency is increasing mainly due to the fact when software will be working on particular file, the entire computer will be most likely frozen for whole week at least due to tens of millions of comparisons and all RAM used, I will mention another example (although i plan in completely different environment - manufactory) where could be the tool used:

Location: Hospital

Situation: Patients - increased number of patients in particular time period due to signs of similar diseases

Filters based on: parameters received in laboratories (blood, urine, etc), where had they been before sign of disease occured, what kind of signs occured, what they were doing, how soon they came to the doctor, location (air, potential dangerous gas, etc) analysis, any different disease particular patients had had before, any remedies they are regularly taking and many many many more - just about anything doctor (or staff in laboratories) would want to know and/or analyze

What I would define in the software: whatever im interested for, e.g. Patients in both gender and any age/nationality, who visited in the hospital in particular time period particular laboratories and went from those laboratores to particular departments/doctors. Obviously I wouldn't define which element of diagnosis would be compared with, e.g. people coming from particular city. This would be expected from the software. If software showed large irregularities per city, I would further dive into the software, run analysis again and this time perhaps set to define I am being interested in elements of air in that city in particular time period - perhaps there were some additions in the air that could risk human's health. Perhaps I would run analysis (I know it sounds crazy but I am further explaining only how I expect the software to work) of police records according to missing people or wanted people due to crime activities - perhaps someone used to randomly set dangerous substances when he/she received status of ''wanted person'' - if report pointed to someone, thats already some result I could pass forward to investigators, etc.

There are billions of possibilities I could define what I am interested for and the software I am looking for would provide the answer based on comparison between values. But of course values (unless logical ones such as mm.yyyy if defining time period) must be given, e.g. full list of people with status "wanted person" must be gathered and located inside parameters otherwise software cannot analyze.

What I would actually expect from the software: Report/answer with graph(s) WHY the disease occured.

I guess by now it's obvious that I don't know of any such analysis program, and I suppose no one else here does either or they would have spoken up by now. But now that we've talked more, I believe such a program is possible in principle. (At first I thought you were asking something impossible.) Have you considered writing it yourself, or hiring someone to write it for you? It would take a lot less time than doing all the analysis without it.

Im not developer/programmer and I don't need this for my personal needs but for job needs so obviously not willing to pay from my own cash. I guess there should be such comparison softwares available... Anyway, do you know any BW expert? Or at least Excel addons' programmer/developer expert? Im not asking where to search for him/her (e.g. freelancer.com) but im asking for direction to specific programming team or contact person.

Don't you think such comparison tool should have existed?

It's never occurred to me that there's a crying need for such an analyzer, but then this isn't the sort of data I normally work with. And I would think that most analysts would expect to look first at the data vectors that seem most likely. But I can see there might be a use sometimes for just throwing a computer at the problem and asking it to calculate all correlations and rank them.

I'm sure you can hire someone to write such a program. In fact I'm sure you could hire someone here on this board, if you could get their attention. But I suspect there'd be a problem with that: This place is provided, after all, for the benefit of newbies who need help from those who won't charge them for their time. As soon as you start talking money, it creates a significant change in the atmosphere, one that the moderators might object to strenuously. And although I'm a capitalist at heart, I don't really blame them; we don't want this place to become overrun with advertisements for skilled services.

Paul, can you advise this guy? Would it be acceptable if NewPuppy were to resort to private messages on this forum? Or can you steer him to another forum that would be more appropriate for this purpose? His need seems legitimate, but I don't think you'd want him pursuing it here, so what's his next move?

I have already said all I think needs to be said - you can't analyse something for which you don't specify parameters. Pure and simple. Software to do that doesn't exist.

When you have stuff like:All I can see is a fundamental lack of understanding as to how any form of statistical analysis is done - whether by people or by computers. Statistical analyses don't come out of the ether - someone has to specify the parameters.

I very much doubt anyone worth their salt here - or on any other forum - would take this on. Still if you want to pursue it, I'd suggest contacting one of the statistical analysis software companies - they'll have the professional statisticians and programmers. Just don't be surprised if you get the same kind of answer as I've already given.

__________________
Cheers,
Paul Edstein
[Fmr MS MVP - Word]

Bob i replied on your private message.