Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[sraymond]

i wanted to be able to graph stuff like (pie or bar-graphs)

-what % do diff genres make-up out of all genres.

- what genres do tracks rated 3 and above fall into and the % spread.

- % of different encoding rates.

- not sure if this can be done with smartlists or not in which case i would use access.

i rate tracks, and to get an over all score for an album i avg the ratings. i was going to make a query that allows me to see avg scores in decimal increments of 0.1 between 2 and 4

wonder if there is an easier way to achieve this ?

I was thinking of learning Scalable Vector Graphics (SVGs), which is a "language" for describing two-dimensional graphics and graphical applications in XML.

Using this technology, I could write an XSLT that transforms an exported MPL into some summary statistics and generates a bar chart, line graph, or pie chart.  IE6 and Firebird (if you get the right build) can both view SVGs.  With Scott.'s MCXMLExport plugin, this would mean you could easily select a playlist or view scheme, choose the configuration for "ChartGenreSummaries" (which I'd provide along with the appropriate XSLT), and click "go" to get the result in a browser window.

Is this something others might find useful?  hit_ny?

Scott-

[hit_ny]

I was thinking of learning Scalable Vector Graphics (SVGs), which is a "language" for describing two-dimensional graphics and graphical applications in XML.

i thought of this intially and looked at the chapter on VML, which is IE's implmentation of SVG. i realised it was interesting but wanted to get up to speed faster w/o having to learn yet another language, ( i did not have your xsl at that time Scott).

Another possibility was a plugin that would do this. But there are many variables and display formats that could be chosen. I finally decided on access as once the data was imported it could be graphed in many ways and the display side of things is handled by access fairly well.

My progress so far :

- what % do diff genres make-up out of all genres.
- what genres do tracks rated 3 and above fall into and the % spread.

These worked quite well in access displaying duration of track vs genre.

- % of different encoding rates.
wil have to do a replace (for the rates that do not fall into CBR defined slots ) as VBR.


The impression i got was they were nice to see, to get an over all idea of what is there. I dont know whether this is something that one needs to do that often.

IF better more immediate use-cases could be defined, i can imagine such an application being more used. The 3 above work on a global basis and do not change as rapidly at least not in my collection

This does not mean the exercise was a waste of time (IMO) There was the possiblity of an XML dump and i was looking at ways/ideas the data could be used.

- To get an over all score for an album i avg the ratings. i was going to make a query that allows me to see avg scores in decimal increments of 0.1 between 2 and 5

this one is taking some time as i have not yet rated everything completetly. I feel this is a useful( more immediate use case) example. IF i ever wanted to know which albums had OVERALL good ratings to recommend to friends.

More examples of usecase like the above need to be found before you have something compelling. Maybe others could suggest ideas.

Using this technology, I could write an XSLT that transforms an exported MPL into some summary statistics and generates a bar chart, line graph, or pie chart.  IE6 and Firebird (if you get the right build) can both view SVGs.  With Scott.'s MCXMLExport plugin, this would mean you could easily select a playlist or view scheme, choose the configuration for "ChartGenreSummaries" (which I'd provide along with the appropriate XSLT), and click "go" to get the result in a browser window.

This sounds interesting, ideally if JRiver decided to incorporate a XML compliant browser ( borrowing from Firebird as JimH implied in an earlier post).

An XML application ( like yours ) could render the graphs in a more seamless way within  MC itself.

In the meantime, it will be easier for ppl not having to bother with access, going from MC to browser immediately.

[sraymond]

I think I might try:

- what # do different genres make up out of all tracks' genres (bar chart)
- what # does year make up out of all tracks' years (bar chart)

and if I get sufficiently motivated:
- a pie chart of genres with percentages.

Pie charts are a bit tough because SVG requires a trig function to calculate the slice.  XSLT has no trig function so I need to find an XSLT library function that does this.

Scott-

[hit_ny]

You might try bar diagrams. It looks like those progress bars and you can indicate different slices within it, maybe with % showing and an arrow pointing to the slice to indicate which category you are charting.

hope this is clear i am not referring to bar charts. Just one bar showing diff slices within it rather than many bars on a chart. This way u can get a few bar diagrams on a page.

Bar diagram might be easier than pie charts and apart from personal preference will show just as much information as a pie chart.

FYI, I'm using 3D exploding pie for charting atm.

[sraymond]

You mean "stacked bar" charts?  That would certainly be easier than a pie chart (I wouldn't have to deal with with lack of a native trig function in XSLT).

Hopefully I'll give this a go this week.

Scott-

[hit_ny]

You mean "stacked bar" charts?

i am referring to horizontal bars. One bar per criteria you want to chart.

e.g.  a bar showing ratings 1-5 and the % they form out of the total.
 __________________________________________________
|   1 (20%)    |   2 (20%)   |   3 (20%)  |    4 (20%)  |    5 (20%)   |
 --------------------------------------------------------------------------------

now the slice size is a function of what % they are compared to the total.

I am not sure what its called. But the above illustration is what i mean.

You can have each bar represent each criteria you want to chart...its more compact and should fit on one page.

[KingSparta]

I think Blues Would Be The Highest Since Blues Is = 0 in ID3v1.1

and most of the time this would mean that the user did not change the genre from the default.

the easyest way would i think to get the FreeDB database and just look thru it (scan it) for the genre. this may give a base graph.

then scaning your own library will give you somthint to compare it with.

[hit_ny]

Eh ?

What u on about King ?

[KingSparta]

Eh ?

What u on about King ?

Say Again All After "Eh"

[hit_ny]

What u on about King ?

What are you talking about ?

[KingSparta]

if your going to look at your genres, you can then compare them what everyone has genres set to by scaning the FreeDB database

so you have 10% rock the rest of the world may have 20% rock and a graph can then be generated what you have and you can compare that to what the rest of the world has.

you may have an artist listed as rock but the rest of the world has that same artist listed as somthing else.

or were you talking about something else?

[sraymond]

e.g.  a bar showing ratings 1-5 and the % they form out of the total.
__________________________________________________
|  1 (20%)    |  2 (20%)  |  3 (20%)  |    4 (20%)  |    5 (20%)  |
--------------------------------------------------------------------------------

Right...  that was what I pictured.  I think the technical name is "stacked bar chart".

Scott-

[hit_ny]

i was talking about an alternative way to pie-charts to display how much of the duration was occupied by x...where x may be rating ( 1...5), genre, encoding rate etc


i'm not sure what one would gain by comparing their % of genres vs what was stored in the freeDB databse ?

[KingSparta]

i'm not sure what one would gain by comparing their % of genres vs what was stored in the freeDB databse ?

well if you rated it rock and they rated it someting else, you would gain the information maybe the rest of the world is living at another slant than you live your life.

you could be wrong about the genre your giving that artist, knowing what the rest of the world thinks may point you in the right direction about fixing the genre.