Hi Mark, Thank you again for being generous with your time. Here is my process: I use a BI reporting tool that is in fact nothing more than an Excel plug-in. For this particular report, I am prompted for a time period which will always be a fiscal quarter. This report returns 5 columns of data: Customer (approximately 250 rows); Rev ordinal rank; Rev $ Amt; GP ordinal rank; GP $ Amt. Thus, I have a 5 col X 250 row matrix of data. Time dimensions are not explicitly captured. Now I have 8 of these matrices. Always, when working with spreadsheets, I just "Babelize" the data into my .json data file. I think you are suggesting that for my measures(properties) I use an array to contain my values, right? RevAmt: ["Q108- $1,000,000",...,"Q409-$1,200,000"], so, if I filter (search facet is probably easiest given number of companies) on Company X then have a list facet for quarters and filter on Q1 how do I get all of the Q1 values for all of the properties? For example, if I have the 4 measures (properties) listed above if I filter on Q1 I would want to see 8 values - 4 values for Q108 and 4 values for Q109. That is where I am lost. Thank you again, Mark, for your time and thoughtfulness. Regards, Kirk
On Aug 26, 8:49 am, mleden <[email protected]> wrote: > Hi Kirk, > > OK, think I might have a clearer picture of your problem now. To put > another way, you *WOULD* model the data the way I suggested but for > the fact that your data source does not supply certain of those values > (for example, YEAR and QUARTER). Do I have that right? Assuming I > do, I would suggest that you pre-process the data before it gets to > Exhibit such that it does indeed include those values. As you point > out, you know implicitly what those values are. Obviously, once > you've done that, you have answered your own question of how to filter > on "time components". If you outline the existing process to generate > the data, I (or others) might be able to offer suggestions on the > specifics of adding the pre-process steps. (From what I can tell, you > run a series of reports. From what system? In what format is the > output? Where does it go from there? How do you feed it to Exhibit? > And so on.) > > Again hoping this moves you forward, > -Mark > > On Aug 25, 8:02 pm, Kirk <[email protected]> wrote: > > > > > Also, in case I am not being clear, my unique identifier (label) is > > the company. Should I perhaps be thinking of using a couple of data > > files and somehow using the combination of the two for my exhibit? > > > -Kirk > > > On Aug 25, 5:58 pm, Kirk <[email protected]> wrote: > > > > Hi Mark, > > > Thank you for your response. I have built several exhibits in the > > > past and was always able to organize my data as you suggest. What I > > > am struggling with here is that my data points are in essence the same > > > each quarter, exactly as you have listed in your response. Quarter > > > and Year are not separate data points at all , but actually integral > > > (unspoken, if you will) parts of what make each measure unique. Take > > > Revenue Amt for instance, I currently have 8 values for it - 1 for > > > each of the past 8 fiscal quarter. When I run my reports I get a > > > revenue number and I know that this is for FY 08 Q1 only because I > > > input those parms when running the report. Which gets me back to my > > > problem, how do I name my columns (properties) so that the implied, > > > but not explicitly captured, time components can be filtered on in > > > facets? Concretely, if I filter on Q1 how do I just return the Revenue > > > Amt values that represent Q1-08 and Q1-09? Add in all my other > > > measure and it complicates this even more. > > > Again,, thank you for taking the time to respond and forgive me if I > > > am being dense. It is just that my measures have the exact same name > > > each time and I just "know" what time period they came from. > > > Best, > > > Kirk > > > > On Aug 25, 4:50 pm, mleden <[email protected]> wrote: > > > > > Hi Kirk, > > > > > Can provide you with some general rules-of-thumb based on my > > > > experience. They actually apply irrespective of the technology, but > > > > from what I can tell with Exhibit, they also hold true with the SIMILE > > > > framework. > > > > > 1. Always decompose your data to its most atomic form. In your case, > > > > I would avoid the temptation to concatenate individual data. If I > > > > understand your requirements correctly, you'd end up with the > > > > following set of "columns": > > > > CUSTOMER > > > > YEAR > > > > QUARTER > > > > REVENUE AMOUNT > > > > REVENUE RANK > > > > GROSS PROFIT AMOUNT > > > > GROSS PROFIT RANK > > > > etc > > > > (The two "RANK" columns are questionable, since I presume that you can > > > > derive the ranking based on the value, but I leave that to you.) > > > > 2. When performance of the application becomes an issue -- and it's > > > > not a question of if but when -- you will need to "tailor" the dataset > > > > accordingly. For example, you could subset so that Exhibit only has > > > > to deal with the last 10 years. Having the source data in a database > > > > makes this a snap, but no matter what there are invariably ways to do > > > > this without impacting the application's functionality. > > > > > HTH, > > > > -Mark > > > > > On Aug 25, 11:19 am, Kirk <[email protected]> wrote: > > > > > > Hi,My apologies in advance for the length of the question - I couldn't > > > > > figure out how to make it shorter. > > > > > > I am trying to create an exhibit of customer data for "higher-ups". I > > > > > have 8 > > > > > quarters of data so far, and will be appending more with each passing > > > > > quarter. The overall population of customers numbers 400-550, with > > > > > around > > > > > 200 of those buying product each and every quarter. > > > > > Each quarter's data is initially captured in excel and, beyond the > > > > > customers, consists of revenue and gp (gross profit) dollar amount > > > > > columns > > > > > and an ordinal ranking for that customer for each of the two > > > > > measures. It > > > > > is a requirement to be able to compare year over year (like quarter) > > > > > changes > > > > > in some, if not all, of the measures. I can either do that as formula > > > > > columns in my spreadsheet or as functions in my exhibit( thoughts?). > > > > > As you > > > > > can see I will basically have the same "properties" for each quarter: > > > > > rev$; > > > > > gp$; rev rank; gp rank; and then any formula columns I may use. > > > > > So, my question is, what is the best way to handle this? If I add the > > > > > values > > > > > as individual columns, I am concerned my dataset may eventually get > > > > > too > > > > > large. If, for each quarter, I add a column (property) like "Q4-09" > > > > > with a > > > > > value for each customer like "1 - 22,000,000 - 3 - 1,750,000" (rank - > > > > > rev > > > > > amount - rank - gp amount), how will the user "massage" the data? i.e. > > > > > compare, say Q4-08 to Q4-09. > > > > > Again, sorry for being so verbose, and sorry if I have been unclear. I > > > > > work > > > > > at a company that is rather backward when it comes to web technologies > > > > > and > > > > > would like to demonstrate the usefulness of the semantic web and the > > > > > tools > > > > > available. As you can see my conundrum is the similarity of each > > > > > quarter's > > > > > data. > > > > > Thank you in advance for any and all help you may offer. > > > > > Best Regards, > > > > > Kirk --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "SIMILE Widgets" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/simile-widgets?hl=en -~----------~----~----~----~------~----~------~--~---
