Forums

MALgraph is quite feature-rich, but maybe there's something we've forgot to add?
If you have any good suggestions, post them here.

Note: don't suggest counting rewatches/rereads - we can't read it, because MAL API doesn't provide such information. This is up to MAL admins to fix, and as things are now, we are simply unable to get rewatch info.

Hopefully I am not sounding rude, but why isn't time calculated of "dropped" titles?

And re-watching is also counted?

And another thing I'd like is :

The color schemes could be changed. Not the bars though (which would be good as well, if you could).

The thing I am talking about the image that is exported. It is transparent (and fine the way it is) but I like to use it in other places than MAL as well and well the text color is not suited for dark stylesheets. So, if there was an option for either dark/light skin for different font colors or just choosing font color (which would be too much, maybe)

I know, you only made this keeping in mind for MAL only, but this was just a suggestion.

Ummmm will there be an option to change color of the graphs ? in older version it was possible to change the colors and it was quiet handy .

Seconded. I'd also like to see the number of dropped shows as many people don't rate what they drop.

Maybe a button to expand "Random friends:" in "all friends:" (to more easily get to their stats)?

I suggested this on the reddit thread:

Some kind of "Dropped Shows" analysis: Most dropped genres, studios.

And a very interesting idea:

Some Staff analysis: Most followed Male/Female Seiyuu also some stats on the best scored Directors, Music Composers, etc.

This was a bug and it should be fixed by now.


It is not counted right now, and I'm not sure how I could implement this since malappinfo.php doesn't seem to return any usable information about rewatch count.


This would greatly lengthen load time (by approximately 5 seconds). For example take Xinil. He has, like, 650 friends, and to list them all, you have to visit 16 separate pages. We might lengthen it to 10 random friends, though, since it can be fetched with one visit to user's profile.

We don't keep track of anime staff, but even if we did...

If someone is concerned about seiyuus, he or she already knows what's up.

Manga would be treated even more marginally since there would be no equivalence for this functionality.

Ummmm will there be an option to change color of the graphs ?

Since this seems to be frequently asked question, I'll try to work on that.
But prepare for this: after you choose colors, colors will lock for 24 hours. So changing them won't do a thing unless the cache expires.

The main reason to do it should be "Discovery".

The ratio of "Concerned" to "Not Concerned" is obviously very low. MALgraph could help a lot by providing this information in a comprehensive way (similar to the prefered studio data) and we would have, for example, people discovering that they love those shows with music composed by the Kajiura Yuki or those directed by Shinbou Akiyuki.

Would it be possible to show how many shows we have watched from each studio?

I know there is a section for our most popular studio in terms of scores given, but I'd also like to see one for most popular studio based on the amount of shows I have seen from each studio.

Type (TV / Movie / OVA / whatever) distribution?

Just something I've been noticing regarding the "List of anime you could've overlooked" section. I'm finding that most of the titles it recommends are of the same series rather than different titles that may be similar.

Ie. If you liked Hidamari Sketch, you may like Hidamari Sketch Specials.

I feel it would be more effective if this section could be used to recommend similar titles rather than just those that are of the same series. Though I'm not sure how feasible it is to implement such a feature.

That's why we named this section like that. These are not user recommendations, but related series (click [?] in the top-right corner of each section to see an explanation).
As of the other (pretty good) ideas: your posts are not ignored, we're watching this thread and considering ups and downs of each. (I always get annoyed when posts with suggestions get ignored on any app's forums, so I wanted to write this.)

This is based not only on the scores, check "Formulas behind the graphs" topic.

First of all, I think it's great that MAL graph is now displaying various statistics in addition to the basic score distributions now. I have some suggestions for improvements:

- I'd like to see a minimum criteria introduced in the calculations. For example, my favoured decade is the 70's, with an average of 9.... because I only watched 1 show from the 70's, and I rated that a 9. Another example is that there are some studios that appear in my most/least favourite studios section even though I've only seen a couple of shows from them. There should be some minimum filter that filters out these results where not enough data has been gathered to produce a meaningful result.

- Could you add a score breakdown by year in addition to decade?

- Favourite studios should be renamed to favourite producers as it works off the producers list, which contains broadcasters, distributers, advertising agencies etc in addition to studios.

Thanks.

One more idea, I would like to throw out is :

Can we have serializations for manga like we have studios for Anime?

Ability to resize the graph, and maybe can we use an image as the background?

Why?

We can implement this, but unless there are more users who'd want it, I'm not likely to implement this. And if it puts too high load on the server, this would be probably first thing we'd disable.

We're working on it.

That is not going to happen. Decades show best mean score, not one`s favorite decade. We could use the same model as we do it for genres and producers, but that would result in whole new diagram, which couldn't be related to global mean score like now.
Secondly, I was taught that discarding outliers is wrong. Current model for genres and studios is in fact an example of balanced quality-to-quantity model (see formulas topic). Unless I am provided formal analysis of other model that yields statistically better results, this is not going change in any way. These studios show up, because you rated highly all the titles.

We could, but this would result in really random and unreadable charts for most users. I'm against this. I'm not sure what fri thinks about this.

Done in one of previous revisions

The older mal graph was much bigger lol, and I think the newer one is a bit to smaller. And what I mean by resizing is the image itself.

And ok

Umm am I missing something, or what happened to your mean score comparison to the global average that the previous MALgraph had? Or am I just being a drop kick, and its staring at me right in front of my face? :S

And what does this mean?

δs = 1.42

You're not missing anything - it's not present here. I already explained it in the shoutbox:


It's standard deviation - http://en.wikipedia.org/wiki/Standard_deviation
I guess we should've used sigma instead of delta; anyway, I'll rename it in the future version.

As for achievement ideas, it's hard to say since there's no list of what is currently available, but so far I haven't seen any demographic-based ones (shounen, shoujo, seinen, josei), which would be nice, I think. Maybe studio-based ones for some of the more well-known ones, too (KyoAni, Madhouse, SHAFT, and JC Staff come to mind first, though there are obviously many more possibilities).