What's suitable option should i choose for interactive report?

[#1]

Hi all,

I'm reaching out for advise. Currently, I'm working on repeated tasks of analyzing data over 1 milion row of data and then showing graphs and reports. Script will automatically analyze data tables, save and closes it. After that, the saved file must be able to adjust certain values through slider box and global box, then update new value to graphs andreports. What's the suitable option that i can build my script on? JMP has provided serveral options but i'm unsure which works:

1. Dashboard
2. Application
3. Interactive Html

Appriciate every advise. Thank you in advance!

[#2 · Accepted solution]

Dashboard and application should both work if users have JMP access. Interactive html might have some problems (depending on reports) to use "slider boxes" to update values.

[#3]

Hi jthi,

My other concern is about file size of the output. Assuming that users have access and my original csv file is up 4-5GB, how large is my dashboard/application saved file?

Besides that, is there any conflict if i analyze multiple files at the same time? I already created a script for my purpose, but when i run it on 2 different tables, adjusting global box on table 1 seems to do the same for table 2 (which i don't want it happen).

Thanks.

[#4]

After you have all required data in the data table, you could save it as compressed file, this will reduce file size most likely a lot. The application/dashboard itself won't be large as it is just JSL script which would open the data table (unless for some reason you want to save raw data to the file which would make it unnecessarily large). If you manage namespaces and data tables well enough there shouldn't be conflicts.

[#5]

Thank you jthi, P_Bartell and Georg for your answer and also insight on how to manage big file in JMP! 

Appriciate all the helps

[#6]

 In addition to the comments from @jthi something to also think about are some of the report consumer/user sharing pathways that are centric to the JMP ecosystem. I'm thinking specifically about JMP Public and JMP Live. Depending on the consumers of the reports and your own sharing specific requirements one or both of these JMP products might play a role in your situation.

[#7]

In addition to answer of others (jsl and dashboard size plays no role in comparison to data size, there are several ways to perform your task), I highly recommend to use compressed jmp file to provide data, in comparison to csv, if the data can be prepared once and used by many. Make sure that preferences are set accordingly or use "Compress File When Saved( 1 )" for the table to make it effective. The filesize will shrink from 60% down to 20% or so, depending on datatypes. Getting several GB's via network is no fun.

Names Default To Here( 1 );

// Script generates random data, stores as compressed jmp file and csv, and compares file size
// random normal = high complexity, size shrink to 60 %
// random integer = low complexity, size shrink to 20 %

form_expr = Expr( Random Normal() );
// form_expr = Expr( Random Integer( 3 ) );

dt = New Table( "RandomData", add rows( 1e5 ), New Column( "col", set each value( form_expr ) ) );

For( i = 1, i < 100, i++,
	dt << New Column( "col", set each value( form_expr ) )
);

jmp_path = "$TEMP\delete_me_test.jmp";
csv_path = "$TEMP\delete_me_test.csv";
dt << save as( jmp_path );
dt << save as( csv_path );

Print( "file size ratio JMP vs CSV: " || Char( Round( File Size( jmp_path ) / File Size( csv_path ) * 100 ) ) || "%" );
Close( dt, NoSave );