Charts and Graphs bugs topic

Main thing: https://www.librarything.com/topic/334487

Some charts happen on multiple pages, so, when leaving bugs, try to be as specific as possible about which page you are on. (URLs are best.)

Under Media on the main Charts & Graphs page-- https://www.librarything.com/stats/aspirit/overview-- a large section over "Paper Book" is unlabeled. My graph shows the arc for paperbacks but doesn't specify what the other section is. (Hardcovers or every paper book that's not a paperback?)

>2 aspirit: The section is too small to display the text; if you hover over the section, it will show you the label for that section. Also, if you click the chart, it will bring up the data in a textual display as well.

>2 aspirit: I believe that unlabeled section is all the books that are classified as "paper book" but that are not classified with one of the subtypes of "paper book".

knerd.knitter (#2):

That's an issue too, but I don't think it's the one aspirit is talking about. I see the same thing (the outermost arc does not cover the full range of "Paper Book":)



I'd been assuming it's Paper Book that doesn't have a lower-level media type (hardcover vs paperback) set.

Incidentally, this is the image I got automatically via the "Save" option under "Share". Notice it's truncated on the bottom.

>5 lorax:

We are working on the graphics. Fundamentally we are now using an external source to make these. It's taking a picture of a screen. We're going to see if we can adjust it.

I think this is a bug, although it might be a case of something just working differently that I expect or want it to. But... I'm having an issue with the "read in" feature for selecting books to view, on all of the pages where it exists. Namely, that it's often not showing all the books I read in the relevant year. For instance, it's showing 6 books in all collections for "read in 2020." I actually have 120 books read in 2020 in all collections. I'm wondering if the problem might involve the fact that I only ever catalog the date finished, never the date started? But if it's looking at the start rather than the end date, presumably it would show zero books.

Some of the earlier years do look right, though, and a few of them are only off by a few books. Hmm, is this maybe something that's still being cataloged?

(That aside, though, I'm digging all the fun new stuff!)

"Paper" and "Paperback" are confusing. I assume "Paper" means not an audiobook or ebook.

>7 bragan: Looking into this now...

>7 bragan: I was about to mention this too. The chart at the bottom of the main stat page shows the correct number of books, but the top toggles don't.

In case it helps, these are the urls to my catalog from each:

Top toggle (incorrect): https://www.librarything.com/catalog/antqueen&collection=-1&filter=read:...
"Read Dates Over Time" (correct): https://www.librarything.com/catalog/antqueen&collection=-1&readdate=202...

Overview > Read Dates Over Time

From the table attached to this graph you can pull up a list of catalog titles for a particular year. However, as soon as you change catalog styles or sort order or attempt to go to the next page in the results, you are taken back to your entire catalog rather than the selection of books you were initially looking at.

>3 knerd.knitter: The section of the graph is certainly not too small for a label. On my graph, it's larger than the book section that's labeled "Ebook".



The unlabel section is also difficult to tell apart from the Paperback section. See the slightly lighter arc that's marked.

>5 lorax: Yet another issue on the same graph.

>12 aspirit: If the labels have to stay horizontal and the word you're trying to fit into that section is a longer one like "Hardcover" then it's not going to fit in the space provided.

>13 shadrach_anki: Oh, yeah, thank you. I see that now. All the text is horizontal and contained inside the arcs.

Then, I don't much like this graph. A large, unlabeled section that can be mistaken as a part of another section is misleading.

>5 lorax: Not all books that are marked as "Paper Book" are also marked as something under that; that's why the arc doesn't go all the way

>12 aspirit: I believe the labels have to be horizontal, so the text for the one will not fit in that section horizontally.

knerd.knitter (#15):

Not all books that are marked as "Paper Book" are also marked as something under that; that's why the arc doesn't go all the way

Isn't that what I said?

>16 lorax: Yes; I'm sorry; I didn't see your text below. I thought you were reporting it as a problem.

The genres and Dewie pie charts legends are cut off both to the left and to the right on my phone. I suspect it is the small screen but they cannot be resized or scrolled so looks ugly.

My publication dates chart shows I have one book published 610-619; when I click to show it in my library it shows Something is Killing the Children Vol. 3, which has a CK original publication date set as 2021-06-15. Looks like the month and day are interfering with something.

And one more issue - when I click on any of the graphs to look at the details, the first character or 2 in the left column in the table at the bottom are missing 009 instead of 2009 in the messages one for example. (On my phone).

While scrolling through the main page on my phone the screen freezes for 5 to 10 seconds or longer before I can continue scrolling. Also happened on the one subpage I tried. Is it because the pages are so long?

>20 AnnieMod: Thanks. We'll take a look at the tables being cut off.

>21 jjwilson61: What phone and model?

I am looking at https://www.librarything.com/stats/AnnieMod/series

Filters: Read in 2021. It pulls only 40 books. Which is way too low - even if you just look at the "Roy Grace" series, it shows only 2 books, I've read more than that this year (and the read dates are set) and it is missing a lot of series.

Where is it pulling the data from if not the read dates?

>24 AnnieMod: Yes. It's a bug we are working to fix right now. Stay tuned.

>23 conceptDawg: TCL 10L

I didn't see this listed above so will add in my note:

If I click through to my books from a graph or table, the first screen is fine. If I change the view for any reason (so I want view B instead of view A) then it loses the filter and goes back to a full listing of whatever collection I was viewing.

As a step by step:
1) Click on the reading dates chart
2) I clicked on Read 2006
3) While looking at the Your Books screen, change style from B to A (or vice versa)

Note that the filter goes away. (First book with filter is one by Piers Anthony. Without Filter it goes to my most recent finish, which is a book by Mary Anne Evans)

Windows 10, Firefox 91.0 (64 bit version)

>27 gilroy: Yep. We're looking at that too.

>27 gilroy: Please indicate which charts you see this on, because I don't think it is happening for all of them. I do know that it happens with reading dates, but if you see it on others, please let us know.

Most of the charts look fine on my phone (a Pixel 2), but the map seems to be Mercator and devotes a truly appalling amount of real estate to Antarctica.

Screenshot:

>29 knerd.knitter: Well, because the reading dates are the only ones that don't seem to have a filter listed at the top of the book grid like the other stats do, so from my test, it's reading dates and the combined Read, Ratings, Reviews chart when you go through a Read link.

On the Cataloging page, Original Publication Date module (https://www.librarything.com/stats/casvelyn/cataloging), when limiting the data to a specific collection, if the collection has ca. 500 books or fewer, the OPD graph combines the years differently than if the collection has ca. 500 books or more.

The issue is that the graph lumps 1800-1999 together if the collection is small. Probably 98% of my reading (and I suspect almost everyone else's) was originally published during these years, so it makes the data really not particularly useful. It also leaves 2000-2019, and 2020- as-is, which adds a weird mix of granularity and {whatever the opposite of granularity is}.

This is also true if you are not limiting the data to one collection, but have fewer than 500 books cataloged in total.

I've been looking at this on my phone, which is not normally something I would do. I've noticed:

1) It's hard to scroll if my finger is not in the right place. For the circle charts I can scroll with my finger pretty much anywhere. For the bar charts and bubble charts my finger has to be on one of the axes. If my finger is in the middle of the chart it won't scroll. Not sure if it's supposed to be this way or if I care since I don't usually use my phone, but it's slightly annoying.

2) Sometimes the map has weird gray lines on it:

They go away or change if I scroll past it and come back

Also agree with >30 lorax: that there's way too much Antarctica. Honestly could just get rid of it all together since there's no such thing as Antarctic nationality.

Under 'collections', it erroneously claims 34 books are not in any collection, but all of them are in 'your library' plus usually one other

https://www.librarything.com/catalog/SandraArdnas&collection=-1&collecti...

Thanks. Will look.

When looking at my list of media, clicking anything below "Spiral bound" is the same - it takes me to the list of spiral bound books (https://www.librarything.com/catalog/norabelle414&collection=-1&formatuser=21)



ETA: If I change to a collection that does not have a spiral bound book in it, it still happens but with the last format under "paper book". It's not specific to spiral bound

>34 SandraArdnas: Fixed. Thanks.

>7 bragan: Fixed. Thanks.

>11 shadrach_anki: From the table attached to this graph you can pull up a list of catalog titles for a particular year. However, as soon as you change catalog styles or sort order or attempt to go to the next page in the results, you are taken back to your entire catalog rather than the selection of books you were initially looking at.

I'm not finding this. Are you still finding it?

>36 norabelle414: I think this has something to do with custom media formats

>19 amanda4242: My publication dates chart shows I have one book published 610-619; when I click to show it in my library it shows Something is Killing the Children Vol. 3, which has a CK original publication date set as 2021-06-15. Looks like the month and day are interfering with something.

The date isn't from the original publication date, but from the publication data in your own catalog. You can see it here, https://www.librarything.com/work/26376803/edit/198874718 .

>24 AnnieMod: Filters: Read in 2021. It pulls only 40 books. Which is way too low - even if you just look at the "Roy Grace" series, it shows only 2 books, I've read more than that this year (and the read dates are set) and it is missing a lot of series.

Fixed. Thanks.

>41 timspalding: Wow! The library that record came from had some bad data. Thanks!

>27 gilroy: If I click through to my books from a graph or table, the first screen is fine. If I change the view for any reason (so I want view B instead of view A) then it loses the filter and goes back to a full listing of whatever collection I was viewing.

This was fixed.

>32 casvelyn: It lumps them by a complex algorithm. I'll have knerd.knitter discuss how it does it.

>36 norabelle414: I think this one for knerd.knitter.

>39 timspalding: It appears to be fixed. It was the same issue that was reported in >27 gilroy: as near as I can tell.

The selection list of collections on the main charts page includes inactive collections. Seems to me it probably should not.

>48 norabelle414:

Thanks. Will check now.

On this page https://www.librarything.com/stats/casvelyn/sitecompare "Blockbusters" is misspelled.

Also I really wish I could see which 4 of my books are considered "esoteric." :)

>48 norabelle414:

Fixed! Thanks.

>50 casvelyn: On this page https://www.librarything.com/stats/casvelyn/sitecompare "Blockbusters" is misspelled.

Fixed that!

Also I really wish I could see which 4 of my books are considered "esoteric." :)

Oh, indeed. I need to make that linkable. I'll look into it.

Overview > Social > Badges
My badges page is blank. Is it just me?

Click on this link. Is it blank?

https://www.librarything.com/stats/Taphophile13/badges

>54 timspalding: Yes, that's blank for me.

>55 Taphophile13:

Can I ask your OS, browser? I'm confused at this.

>56 timspalding:

Yeah, I'm quite confused. I even went in as you in two browsers. Odd.

>56 timspalding: I'm confused too.
Win 10, Chrome 92.0.4515.131 (Official Build). Also blank on Firefox 89.0.2.

The following books should show up when I select "read in 2013" (https://www.librarything.com/catalog/norabelle414&collection=-1&filter=read:2013%7Cno_reload:1) but do not:

https://www.librarything.com/work/14149720/details/103715317
https://www.librarything.com/work/10746533/details/103993039
https://www.librarything.com/work/22436/details/103997066
https://www.librarything.com/work/10522105/details/104534535

I can't tell what they have in common. They were all entered on the same day as the "finished" date, but that's also true of several books that do show up on the list. None of them have multiple reading dates.

>58 Taphophile13:

Okay. conceptdawg, you want this one?

>59 norabelle414:

Investigating.

>59 norabelle414:
I have another long list for 2014 (https://www.librarything.com/catalog/norabelle414&collection=-1&filter=read:2014%7Cno_reload:1). All of the books (and the ones above) are:
1) manual entry (almost all of mine are)
2) entry date is on or after finished date (but they're not the only ones)
3) I got them from the public library (but they're not the only ones). This might only be related because sometimes I don't catalog library books until I finish reading them, whereas books I own I catalog as soon as I acquire them.

https://www.librarything.com/work/15067844/details/114608638
https://www.librarything.com/work/13542748/details/113408594
https://www.librarything.com/work/12803174/details/113408559
https://www.librarything.com/work/12336459/details/113408503
https://www.librarything.com/work/12877924/details/109063192
https://www.librarything.com/work/11116632/details/108886560
https://www.librarything.com/work/10577436/details/108770392
https://www.librarything.com/work/9438501/details/108489965
https://www.librarything.com/work/2369220/details/105360488

>61 norabelle414:

I found the problem, which is in the data, not the new code. Fixing.

I noticed something strange in my Genre charts
https://www.librarything.com/stats/HeathMochaFrost/genre

Specifically in "Entry Dates By Genre." Right now, the top 3 labels (along left margin) are Art & Design, Business, and Christian Fiction. There are good-sized dots BETWEEN Arts & Design and Business, indicating Biography & Memoir...which isn't listed in the chart. If I hover over the first larger dot, it'll tell me I entered 17 Biography in July 2007. The total Arts & Design for all collections, all time, is currently six, and Business is only five. Total Biography & Memoir is 145.

I noticed the "missing" labels when I turned on all the Genres in my Preferences, just to see if that impacted any of the data, and I was surprised to see Food & Cooking in that list. I literally have two food books. All the dots in that area were actually for Fiction & Literature, which had disappeared from the list. My total Fic & Lit is 1,974. In the example above, I'd removed Food & Cooking, and Fic & Lit was back in its spot, but it's followed by General Nonfiction; General Fiction (I have 721 of them) isn't in the list.

The dots are an interesting way to visually represent the data, but if there are gaps in the labels, making it appear that we have lots of books in a Genre where we have almost none (or vice versa), this chart in particular is more abstract art than a way to present and visualize our data.

>63 HeathMochaFrost:

For knerd.knitter, I think.

>59 norabelle414: >61 norabelle414:

Fixed. Thanks!

This might be deliberate, but I do find it odd that the Nationality map uses different projections on my laptops (plate carrée) and phone (Mercator).

>58 Taphophile13:
Do you have an ad blocker? My badges list was blank until I turned mine off.

Some of the graphs omit information that is unknown/not set on the main page view, but include it in the zoomed in view -- i.e. the graphs for living/dead authors and author gender.

Is this intentional? It seems a bit odd to see one graph, and then click on the expanded view and be presented with a graph that looks very different.

In Nationalities (I use the Dutch LT), Reggio Emilia, Noord-Italië is part of the UK, instead of Italy.
I could change it on Ludovico Ariosto's page, where it comes from, but then this region is still misplaced.

When I look at the author gender chart, it shows "female" and "Female" as different genders (same with "male" and "Male"). I guess I could go in and change the CK on the upper case ones, but I assumed the system automatically recognized that they were the same thing. Capitals shouldn't make a difference when it comes to gender (it might make a difference elsewhere).

There is some code in the top right of my "Vous et nul autre" page, where "Help" should be.
It is not on the English LT (.com), only in translated sites (I checked Dutch, German and French).

Screenshot https://www.librarything.nl/stats/FAMeulstee/vousetnulautre


Close up:

My book cover chart seems very off. It claims the whole doughnut/donut is member-uploaded covers (which I think is true), but the number stated is 36. Where does that number come from? I have 821 books catalogued and can't really see what these 36 books have in common. The vast majority of my covers are images taken from elsewhere online, and I've never uploaded a photo myself to LT.

>72 PawsforThought: Which filter are you using?

>73 anglemark: I just clicked on the link in the chart. The doughnut, when expanded, has (from the centre outwards) a complete circle with "All", a complete circle with "Member uploaded cover", a complete circle with "Member-uploaded cover, best-guess for ISBN" - all with 36 books for each - and the outermost circle is split between "Low-quality: Member-uploaded cover, best-guess for ISBN 1" (i book) and "High-quality: Member-uploaded cover, best-guess for ISBN" (35 books).
When I click to go to the catalogue, it's set to All collection, but the filter is "coverstats::isbncustom". I don't understand why only those books are included and where the other are.

>74 PawsforThought: Yes, I see the same when looking at your stats. Weird.

Tag mirror seems to display the same number of tags no matter what I select, if I have applied the "Read in 2021" filter: https://www.librarything.com/stats/Stevil2001/tagmirror (I did not experiment with other filters.)

>75 anglemark: I'm confounded.

>76 Stevil2001: What exactly are you trying to select?

When I open up a graph, the data in the table below cuts off the first character in the left hand column and I can't scroll across to see the full word.

eg. For Your Books Over Time, that column says: une 2006, uly 2006, ugust 2006, etc.

I'm using Android 11 and the latest version of Chrome.

A couple things:

On this page: https://www.librarything.com/stats/casvelyn/genre in the Fiction vs. Nonfiction module, can we set the genre for the Not Set books? I can't find how to do that in the Genre field in my catalog. Also I think I see one place where the algorithm is getting confused on this. Most of my books that are Not Set are Children's Books. Because Children's Books don't usually get classed as a genre besides Children's Books, the system doesn't know where they fall in the Fiction vs. Nonfiction divide. (Example: the American Girl books are Historical Fiction, but the genre system calls them Children's Books, so they get left out of the Fiction class.)

On this page: https://www.librarything.com/stats/casvelyn/ratingsreviews in the Read By Year module, it has "half years" that don't really make sense there. Plus they are written as numbers rather than dates. So it goes 2018, 2,018.5, 2019, 2,019.5 etc.

ETA: The Publication Dates module here https://www.librarything.com/stats/casvelyn/cataloging doesn't display every decade. I've read at least one book where my copy* was published in 1910. I understand truncating it on the Overview page, but it should show every decade represented in my catalog on the Cataloging page.

* https://www.librarything.com/catalog/casvelyn&deepsearch=The+Window+at+the+W...

>77 PawsforThought: Different numbers of tags.

>80 Stevil2001: Yeah, but what are you actually doing? You select "Read in 2021" and then? Click on a tag to see how many books are tagged with that? Or something else?

>67 ErlendSkjelten: Yes! That's what is happening. Strange, that page used to work with the blocker on. Thank you.

Under Nationality, Hongyang is shown as one of my authors from China.

What I've actually have is a book by Yang Hong, who is only combined with Hongyang because of LT technical limitations.

Would it be possible, in the case of split authors, to link, and show the correct name for, the split rather than the combined author?

"Fiction vs Nonfiction" identifies books that has the genre "general nonfiction" as "not set"

Eg
https://www.librarything.com/work/5412931/book/143315406
https://www.librarything.com/work/2867206/book/151822198
https://www.librarything.com/work/10855829/book/68985886

>81 PawsforThought: I click on the toggle for the number of tags you can see!

>84 paulmdh: "Nonfiction" and "General Nonfiction" are different genres.

Regarding the Fiction vs. Nonfiction chart: this only displays what books have the "Nonfiction" genre or the "Fiction and Literature" genre as those are considered "super-genres." If neither of them are on a book, it will show "Not Set"; if for some reason both are listed on it, then there is a "Both" category that will be displayed.

>85 Stevil2001: Are you sure you have used more than 100 tags this year? If not, then there would be a difference, right?

>86 knerd.knitter: I understand that this is intended behavior, but I think it's counter-intuitive to what I at least would expect from this sort of statistic. I have books that are properly categorized into sub-genres such as Mystery, Horror, Science Fiction, Historical Fiction, etc., but they appear in Not Set if they don't also have the Fiction and Literature genre set).

Among those, Historical Fiction, General Fiction, Recent Fiction, Science Fiction — genres with "fiction" in the name — at a minimum need to be included in this chart for it to have any value whatsoever to me. I thought the whole point of having fiction sub-genres was so that not everything would be lumped into one place — am I wrong about that?

>88 rosalita: I agree. Most of my reading is fiction, but most of it is not Literature.

Maybe this has been pointed out, but on the Books->Measurements section, there is a button to switch from metric to imperial. When I click that, the height of my book stack changes. I would have expected units to change (there are no units on the graph anyway), but not the height of the stack. Is this then showing books that are reported in metric instead of imperial? Weird. I mean the value is there, all books should be combined and then reported in either metic or imperial. At least that is what I would expect.

Also, weird that height is given in feet (3.5?? I have more books than that!), but volume is given in cubic meters (metric), weight in kilograms (metric), and pages (how is this calculated?, I guess it is not total pages? which is unitless?) in meters also metric. Couldn't one system metric or imperial be used for all of those?

>32 casvelyn: I have made some tweaks to how the original publication dates get grouped; you should see them in a few minutes

The algorithm is a little complicated, but basically for both original publication date and publication date, first it groups everything into decades; then original publication date attempts to group into 200 year periods (only if there are more than 200 dates represented in the initial data); then for both, it attempts to group into centuries; then it attempts to group into half-centuries; then it attempts scores (20-year periods). The Original Publication Date grouping is a little less stringent because there's a potential for so many different periods to be represented. Most people don't have books whose editions were published BC, but many will have ones with original publication dates of BC. It uses a count cutoff so it doesn't group decades with too many books (15 for Original Publication Date and 10 for Publication Date). We are always tweaking with these things to try to get them to display well, but everyone's data is different, so it's tricky.

>90 al.vick: I'm not sure what you mean about the height of the book stack changing. Does the graph look different? The height should change from feet to meters, but just converted (if the conversion seems wrong, let me know).

When you say the height is given in feet, volume in metric, etc., you mean the ones at the top, correct? Those are just a random sampling of the data that's at the bottom of the page (where the pages length is explained as the length if all the pages in all the books were laid end-to-end). Those stats change when the page is reloaded too, just to give you an idea of what is in the bottom section of the screen.

>36 norabelle414: That was a bug; thank you for finding it. I'm pushing a fix right now, so you should see it in a few minutes.

In metric, my book stack is shorter than a hobbit, in imperial it is the same, huh. I just assumed that was erroneous since my book stack is should be ***much*** higher.

but why feet (imperial) and other in metric?

>63 HeathMochaFrost: This is a function of the library we're using for the charts. If there isn't enough room on the y-axis to show all the categories, it will leave some out, but when you hover over the bubbles in between the lines you should see the right label. We can't do much about this other than making the chart bigger, but we don't want to make any chart too big. For now, you can see the data on the hover.

>66 AndreasJ: This is deliberate. The Mercator projection was a little more square and seemed to be a little better on a narrow screen.

oh, I see the height is based on the collection and it got stuck on something else. Still strange units. now I see weight as in units of adult badgers. Funny but not really helpful.

>68 spiphany: That is intentional so that the "actual" data is shown a little more clearly on the smaller chart.

>70 PawsforThought: Those are treated as separate to encourage people to fix the data.

>87 PawsforThought: This is correct. If there aren't more than 100 tags in the results, you won't see any difference between the tag count options

>74 PawsforThought: There are currently issues with the covers' chart; stay tuned for more information.

>87 PawsforThought:, >101 knerd.knitter:

It's tag mirror, though. Surely more than 100 tags have been applied to the 78 books I've logged so far this year.

>91 knerd.knitter: Thanks! It does work better now for the most part.

The places where it's still weird also make it hard to compare data because the groupings aren't consistent. It would make more sense to me to make static date ranges, so the data can be used more easily. For example, if I filter to Books Read in 2020 or Books Read in 2019, the 1800s and 1900s are divided into 50-year increments. For Books Read in 2018, the 1900s are divided into 50-year increments, but the 1800s are divided by decade. (And logically that should be reversed, because I read 14 books from the 1900s but only 2 from the 1800s, so the added nuance would be more beneficial in the century with more books.)

>79 casvelyn:
On this page: https://www.librarything.com/stats/casvelyn/genre in the Fiction vs. Nonfiction module, can we set the genre for the Not Set books? I can't find how to do that in the Genre field in my catalog.

You should be able to double-click on the Genre field in your catalog and select genres for the book

On this page: https://www.librarything.com/stats/casvelyn/ratingsreviews in the Read By Year module, it has "half years" that don't really make sense there. Plus they are written as numbers rather than dates. So it goes 2018, 2,018.5, 2019, 2,019.5 etc.

This was a bug; I have fixed it.

ETA: The Publication Dates module here https://www.librarything.com/stats/casvelyn/cataloging doesn't display every decade. I've read at least one book where my copy* was published in 1910. I understand truncating it on the Overview page, but it should show every decade represented in my catalog on the Cataloging page.

As explained above in >91 knerd.knitter:, there is an algorithm that is used to determine how to group the decades. It would be impossible to display every decade for some people’s libraries, so we have to do our best to group the data where we can

>104 casvelyn: Sorry, I see you saw my publication date explanation; I was posting while you were, and I missed your response. : )

>84 paulmdh: Thanks. The algorithms are different, but nothing that makes "general nonfiction" should fail to be in the (far larger) non-fiction category. I will look at it.

Phone vs. laptop UI difference:
The "Read in" dropdown is missing on the phone.

As a result, if you select a "Read in" filter on the laptop and then open the page on the phone, the selection holds (it shows 147 books selected for me just now) but there is no indication anywhere on the screen what these books are. Maybe "Custom" needs to be highlighted? Or add the missing filter?

>106 knerd.knitter: LOL the dreaded cross-post!

>105 knerd.knitter: Fiction vs. Nonfiction: Sorry, I wasn't being very clear there. (Never post in web forums when you're in a hurry!) The issue is better described in >84 paulmdh: and >88 rosalita:

"Half years": Cool, thanks!

Pub dates, original or otherwise: Yeah, I do understand that everyone has wildly different data. But if the graph isn't consistent, it's not useful. If data is omitted, the data isn't useful. For the Pub Date (of my own edition) graph, the data stops at 1940. Granted the vast majority of my editions were published between 1940 and the present. But if there could be at least a bar for "Everything pre-1940" that would be more useful and more accurate. Or have everything prior to "year where splitting by decade no longer makes sense per the algorithm" divided by 50 or 100 years.

If data is omitted, the data isn't useful. For the Pub Date (of my own edition) graph, the data stops at 1940.

You're saying it's omitting years, not just combining them?

>23 conceptDawg: It seems likely that my problem was caused by a problem I saw someone else report which had to do with not being able to scroll using the graph surfaces. That seems to be fixed except that I still can't scroll using the last graph on the main page, Talk Messages

>110 timspalding: On this page: https://www.librarything.com/stats/casvelyn/cataloging in the Publication Dates section, it has every year 1940-present broken up by decade. (ETA: Not strictly true, 1900-1949 is lumped. :( ) Then the last bar in the graph is No Date.

But I have books where the edition I read was published before 1940. The example I have off the top of my head is The Window at the White Cat. My copy was published in 1910. But there's no bar on the graph that includes 1910, even if I click on the graph to expand it.

Oddly, I have the same book on my other account, and on that account's graph all the decades show up correctly. (I think. I haven't checked all the books.)

>108 AnnieMod: The Read In filter is removed on the phone-sized screens because there just isn't room for it, and you can get to the filter via Custom. But I'll see about correctly highliting Custom in that circumstance (the Read In is highlighted because it's a "Read In" selection...it's being "smart" but not smart enough).

>112 casvelyn: on that page I see 1900-1949 as the bottom label before "No Date"

>113 conceptDawg: Figured this is the reason for not having it there but the lack of highlighting threw me off a bit last night so figured I should report it :)

>114 knerd.knitter: Oh good grief, don't mind me, I can't read! And I even pointed that out to myself in >112 casvelyn: and still missed it!

I was reading the last bar before No Date as 1940-1949. Which is not, of course, what it actually says.

Sorry for all the trouble! I'm going to go surrender my library card now, for lack of competence in reading... :(

>116 casvelyn: Don't feel bad! There's a lot going on in here! Thank you for all your feedback; it's been helpful.

Very minor, but for the height of the book stack, I suspect the quoted height of Notre Dame is no longer accurate, as it appears to be the pre-fire height including the spire.

All of the quoted heights are of course implausibly precise, but I've done that rant before to no avail. Even if the site you snagged them from is unreliable in that regard, the least you could do is round for display.

>118 lorax: The precision on the stack height chart wasn't really accurate because lots of the heights are listed as whole numbers (https://wiki.librarything.com/index.php/HelpThing:Charts_and_Graphs:Overview), but I've made a change to remove the 0s after the decimal point so they don't look like they're indicating that level of precision. (Change should be available shortly.)

Stats change-related is that on Local, all the venues are starred. (Related to double-checkmark in talk bug? https://www.librarything.com/profile/amanda4242 ) Likewise in Local Helpers log -- all starred.

I don't care whether the "we should show 450 feet as 450.00 feet" decision was made in the source data or on your end, I'm just glad you aren't doing it any more!

>100 knerd.knitter: I don’t think that answers my question. Why is capital letters a different thing? Why is one right and one wrong? “Female” and “female” don’t mean different things. And which is the “correct on, according to LT? And why is that one more right than the other one?

>122 PawsforThought: I have these thoughts about "non-binary" and "nonbinary" showing as different genders on the graph. Of course the entries in CK are different for enby authors-- different parts of the world disagree about the hyphen-- but I don't understand why the graph needs to separate the two. Neither term is wrong and needs to be fixed, and both mean the same thing for the purpose of the image.

>122 PawsforThought: And if Tim thinks that one way is wrong why doesn't he just fix the data in the data base instead of relying on users to do it?

>120 davidgn:
Thanks. I'm looking into those.

>122 PawsforThought:
>123 aspirit:
>124 jjwilson61:
I don't think we have a stance on whether one is more right than the other. Just that there are different values in that data and showing both encourages people to "correct" that.

>122 PawsforThought: And which is the “correct on, according to LT?

The examples shown when you go to edit show lowercase, so I'm assuming that's the preferred LT format. I do agree that it shouldn't differentiate between cases.

We try to let LT members control as much of the data as possible, without changing it or otherwise altering it. People have reasons for putting data in particular formats. Usually the data settles on an agreed-upon format, eventually.

This feature just happens to be shining a light on some data that hasn't reached that point.

>122 PawsforThought: >100 knerd.knitter: knerd.knitter: I don’t think that answers my question. Why is capital letters a different thing? Why is one right and one wrong? “Female” and “female” don’t mean different things. And which is the “correct on, according to LT? And why is that one more right than the other one?

Mere upper/lower is tricky I agree, but there are also transmasculine, trans masculine, trans-masculine and so forth. Do the distinctions matter? They may to some. Others will insist that the "trans" has no part in descriptions of gender, because a trans man is a man, a trans woman is a woman. Members wanted free text, and have used it freely. We urge members to achieve consensus and use the terms the authors themselves use. We have established multiple tools for finding variants and changing them, if necessary. But we're not going to establish authority control.

>129 timspalding: So because other terms might not work if they’re combined, terms where the only difference is capital letters shouldn’t be touched? I’m sorry, but that’s a stupid rule. Or non-rule. This isn’t about establishing authority, it’s about having like be like.

Capital letters isn’t “tricky”, it’s the *exact same word*.

I agree. Capitalization isn't the same issue as whether a hyphen is inserted.

I used to know someone who claimed to be deaf but not Deaf, so I wouldn't be shocked to learn that someone somewhere insists that they're female but not Female, or vice versa.

(The distinction he made was that deaf = unable to hear, but Deaf = member of a subculture for deaf people.)

>131 jjwilson61:

I'm not building a layer of combination logic for capitalization-combination. Because we can't just lower-case everything. I'd have to compare versions, pick the most common, etc. Nope. Members do this data.

On another topic a member writes "Capitalisation seems to be a problem, I have thousands under "male" but also one entry showing separately as "Male". Roger Marshall"

There's a solution: Change one author. That is simply a better solution.

So not a bug, technically. But if you get a relatively small number of books from a relatively large number of places, the From Where graph on the Cataloging section looks... interesting. To be fair, it looks fine when you expand the module.

This is this page: https://www.librarything.com/stats/casvelyn/cataloging, filtered to Read in 2017. I read 47 books from 31 libraries/bookstores.

>133 timspalding: It's not that complicated. One solution would be to just use the JavaScript toLowerCase () function on the field when creating the chart. That doesn't change the underlying data but maybe that's ok.

The "Nationality" rubric is all kinds of ignorant and tendentious shitshow but as I value my sanity I won't go into details... suffice it to point out an asymmetry that ought to be easily rectified: right now, the entry USSR "maps to" Russian Federation, but the entry Yugoslavia maps to... Serbia. Yugoslavia was a federation of six republics and two autonomous regions. Kindly change the mapping so that those assigned to the "Yugoslavia" rubric map to the Yugoslav Federation.

>136 LolaWalser: There is no Yugoslav Federation on the map. The CK should specify in 'country for map' field what existing country to map to, so I assume entering correct data there should rectify the problems. Edit: in case it's unclear - on author page. Individual authors are assigned a country on existing current map of the world.

>137 SandraArdnas:

OK, thanks, just to confirm I understand--the CK on author pages, or...?

>137 SandraArdnas: Yes, I realized belatedly that might not be clear and edited

Thanks, I'll deal with it. (but I'll try not to look at the resulting "map"... my blood pressure's high enough :))

Ah, no, this garbage won't work, the Deciders have Decided things must map to "existing" countries. God save us from the impure-blooded mongrels and travellers! Everybody tick ONE box, only!

Which still doesn't explain at all how the fuck does someone like Kiš, born in Montenegro (Kingdom of Yugoslavia), of Hungarian, Jewish and Montenegrin ethnicities, dead in Paris, end up mapping to "Serbia".

There really should be some warning appended, to the effect what useless garbage this rubric is for anyone with a genealogy and history more complicated than a lab rat's.

>141 LolaWalser: I'm not sure what you want it to do. It can't use multiple maps, nor can it map to more than a single country. FWIW, I consider Kiš part of Serbian literature, but perhaps my high school education was not as good as I think if he should really be mapped somewhere else. Either way, whatever is the most appropriate existing country should be in the country for map and it will map there. For anything more complex, one has to read a bio, not expect it on a stat graph. I forgot to add, if a country for map is not specified and the nationality field has more than a single entry, the last one is used AFAIK.

I think a thread for discussing cases where the current mapping is dubious is a good idea. Is Kundera supposed to be French, for instance?

I was thinking of a similar case yesterday - an author born and bred in Soviet Russia but moved to Sweden in early adulthood where most (all?) of her books have been written. Is there a consensus how to treat such cases for the map? I notice Asimov is mapped as Russian, which I wouldn’t have expected.

>141 LolaWalser: >136 LolaWalser: The "Nationality" rubric is all kinds of ignorant and tendentious shitshow but as I value my sanity I won't go into details... suffice it to point out an asymmetry that ought to be easily rectified: right now, the entry USSR "maps to" Russian Federation, but the entry Yugoslavia maps to... Serbia. Yugoslavia was a federation of six republics and two autonomous regions. Kindly change the mapping so that those assigned to the "Yugoslavia" rubric map to the Yugoslav Federation.

The Yugoslav Federation does not exist anymore. We can only map to countries that exist now. This is how the mapping API on Plotly works. It has current countries. It doesn't have a giant list of past countries, which would, obviously, overlap in terrible, confusing ways.

We can consider mapping to dots on the map. But we have to choose either dots or currently-existing countries.

Which still doesn't explain at all how the fuck does someone like Kiš, born in Montenegro (Kingdom of Yugoslavia), of Hungarian, Jewish and Montenegrin ethnicities, dead in Paris, end up mapping to "Serbia".

You can change him to mapping to Montenegro.

I notice Asimov is mapped as Russian, which I wouldn’t have expected.

Change him to the US, if you like. I'd do that.

>144 timspalding:

I could do that, obviously. But before I start changing stuff it'd be nice to know if there's a convention or consensus for handling this sort of cases.

>144 timspalding: A question I think is relevant: does it also draw on the "Country for the map" field? I've often tried to solve such issues by filling that out.

>146 Nicole_VanK: yes

>147 timspalding: :) Thank you

If I try to chose a collection in the dropdown on the author gender page, most of my collections are hidden behind the chart.

>144 timspalding: It's a problem with lots of people I know trying to limit them to 'belonging' to a single country. For example, I know a Canadian woman of Japanese heritage who married a German. Their children were born in Austria (no right to Austrian citizenship) and have English as their first language. Are you going to call them Japanese because they 'look Asian'?

>149 MarthaJeanne: How would you propose to solve the problem, assuming that we do want the Nationalities map?

Author nationalities were always going to be a problem -- due to geopolitical borders changing but also due to many authors having connections with more than one place. This has come up again and again in previous discussions of the matter.

One option might be to allow authors to count for more than one country on the statistics page. It is possible to add more than one nationality in CK, but it appears that the statistics only count the first item entered (or the "country for map", if entered). This wouldn't fix the problem of how to assign map locations for people from countries that no longer exist, but it would at least allow multiple national affiliations to be acknowledged in the more detailed country list.

Regarding the nationality issue: if you set “Poland” as nationality, it autocorrects to “Russia”. I meant to write Polish, and when I corrected it there was no autocorrection. Might want to look into that issue, though.

>151 spiphany:
Adding to this: the detailed country list for author nationality on the stats page is rather frustrating because -- at least on a non-English LT site -- it apparently pulls the country names from various language versions of CK, so most of my German authors are listed under "Deutschland" with a handful under "Germany".

It isn't clear to me which field ("nationality" vs. "country (for map)") is being given priority, or which language entry in CK is being used -- e.g. Walcott is listed as "Sante-Lucie" (French CK) even though I'm on the German version of LT and both the German and English CK have "Santa Lucia". So I don't know how to fix the CK information so that at least those authors who are undisputably from the same country appear together.

Can you put in a range for the "Read In" filter? I put in "2003-2021" but a lot of the authors that come up shouldn't be appearing, because they are on books I have not recorded reading dates for (either they predate my tracking, or I haven't read them yet).

Four of my books are listed as having No Date for Original Publication Data. However, two of them (Heidi and Gulliver's Travels) have a year in the Original Publication Date field in CK. I've deleted and re-added the OPD, but the books are still listed as No Date. Is there something else I need to edit?

Also, Lassie has an OPD of 1940, but the graph places it in AD 190-199. I guess dog stories truly are an immortal genre. :)

I don't mind fixing these (I love fussing with data) but I'm not sure what field I need to use to correct them.

>155 casvelyn: At least for the Lassie book, it seems to be a case of another CK language causing trouble. The German CK has OPD as “196?” so that’s probably the cause of the problem.
Haven’t checked the others but could be the same or similar.

>156 PawsforThought: Thanks! That makes sense.

Sadly, I fixed the German CK for Lassie and it still says 190-199. I also noticed that The Bears of Blue River is included in 1600-1699 even though it was published in 1901 and there's no OPD in the CK for any language. So the date must be drawn from somewhere else, at least on some books.

>149 MarthaJeanne: What OS/browser/device are you using when the dropdown goes behind the chart?

157 I still see "196? ?" in the German CK for Lassie it's fixed now

>158 knerd.knitter: My iPad, Safari

>152 PawsforThought: Regarding the nationality issue: if you set “Poland” as nationality, it autocorrects to “Russia”. I meant to write Polish, and when I corrected it there was no autocorrection. Might want to look into that issue, though.

What do you mean by "autocorrects" here? I'm wondering if this is your browser, or the helping data, or if it resolves in the end to Russia.

Vous et Nul Autre is not displaying correctly - the totals are incorrect and moving to a second page of results sends you to the wrong page.

If I look at Shared by members None I'm told I have 198 books that only I have - but this seems to be the number on the first page of results - and there are 28 pages of results listed.
And if I click on the second page of results then I'm taken to the second page of the results where I'm sharing with one member, rather than no members.

I can only actually look at all 5 pages of books where I share with one member by using this method. If I look at the first page of results where I share with one member and then try to look at the second page of results then I end up on the second page of results where I'm sharing with two members...
Where I share with one member I'm told I have 169 books on page 1, 155 on page 2, 120 on page 3, 122 on page 4 and 126 on page 5

In a similar way I can look at the 3 pages of books I share with two members. I'm told I have 99 on page 1, 85 on page 2 and 52 on page 3

>161 timspalding: I mean, I wrote Poland, pressed enter and after a millisecond of showing as Poland it turned to Russia.

I haven't tried entering any new authors as Polish, but the ones already had were showing up correctly. (Stanislaw Lem was the first I checked).

>164 lorax: I was correcting Isaac Bashevis Singer, who had USA (correct) and Russia (egregiously wrong) as nationalities.
I erased Russia and wrote Poland. Clicked enter and realised I should have written Polish but it had corrected back to Russia.

>157 casvelyn: Sadly, I fixed the German CK for Lassie and it still says 190-199.

You can only edit other CK fields if you're logged into the corresponding site. So, you'd need to log into https://www.librarything.de to fix the data. I'll go in and do that now.

ETA: and fixed!

>166 kristilabrie: You can only edit other CK fields if you're logged into the corresponding site

Is this new/intentional? I'm used to being able to click on the list of languages at the top of the CK to edit language information for a site other than the one I'm on and I was under the impression that is how it is intended to work.

However, there is or was a bug connected with this -- at some point (in the last year or so I think, around the time either the series or gender field was updated) changes made to CK in other languages stopped saving. It was reported here: https://www.librarything.com/topic/327497
I had this issue earlier today but later on made other changes that do seem to have saved, so I don't know exactly what is going on (I did log on at some point with my second account on the other language site, but I used the tab at the top of CK and didn't switch sites when editing.)

The stats related to stack height/shelf width are way off for my collection. It appears that approximately 1 in 5 books (>600) have the length and thickness measurements reversed (I'm guessing these refer to width of cover and width of spine). Presumably the number of books where the latter is bigger than the former is vanishingly small so that's quite a large data problem but one that should be relatively easy to catch. One of these claims to be 214 inches thick, around 17 feet, suspicious even for a set of encyclopedia, never mind a single volume.

>168 tallpaul: This is a known (and common) problem with book measurements. I don’t know if the cause is known (bad data from source, I’d imagine, but which source/s?)

>96 knerd.knitter: Update on this: I don't know if some settings or code has been changed or tweaked, but the chart "Entry Date by Genre" that's on my Overview page includes the most common genres in my library, and the labels in the left margin don't change when I add and remove things in my Genre Preferences settings. It's only in the Genre Charts & Graphs page (or sub-page, or sub-section) where the chart with the same name and basically the same layout has labels in the left margin that change when I add or remove things in my Genre Settings. (And yes, it's still doing that this evening.)

If a change was made so the Overview version of the "Entry Date by Genre" chart lists the genres that are most common in the user's catalog, thank you for doing that. If it was already like that and I somehow didn't notice, my apologies, and I'll plan to just refer to that one in future and not look at the one on the Genre Charts & Graphs page.

Just one more question: you wrote, This is a function of the library we're using for the charts. This made me wonder: What library are you using?

Thanks for all the hard work, and trying to address so many issues and questions! :-)

>52 timspalding: I'm also eager to see which of my books are "esoteric" -- 33 in my "your library" collection and 90 for all collections. I know you're squashing bugs and answering questions, but it would be super-cool if the Book Popularity info could be linkable, so please keep it on your radar. Thanks!!

>171 HeathMochaFrost:

Thanks. On the list!

Where is Original Publication Date being calculated from?

All of my books have a Publication Date but less than 50% have been given an Original Publication Date.

Checking some of those listed it's not taken from Original Publicaion Date in Common Knowledge.

And some are completely wrong - just a small sample:
https://www.librarything.com/work/7568606 was written in 2010 about a current event - but the Original Publication Date is listed in the range 1800-1899
https://www.librarything.com/work/13940631, a novel based on events in the 1860s, was written in 1974 but the Original Publication Date is listed in the range 1800-1899
https://www.librarything.com/work/191738 was written in 1937, but the Original Publication Date is listed in the range 1950-1959
https://www.librarything.com/work/5104236 is a 2012 copy of a book first published in 1923, but the Original Publication Date is listed in the range 1960-1969
https://www.librarything.com/work/13334414 is a 1969 copy of a book first published in 1868, but the Original Publication Date is listed in the range 1960-1969
https://www.librarything.com/work/5188079 was published in 1951, but the Original Publication Date is listed in the range 1960-1969
https://www.librarything.com/work/1067299 was published in 1906, but the Original Publication Date is listed in the range 1960-1969

At "Vous et nul autre" pages in the non-English sites there is still a bad link "https://wiki.librarything.com/index.php/HelpThing:Charts%20and%20Graphs:%3Cspan%20id=" at the top right, where "Help" should be.

>173 newcrossbooks: I looked at just the Pankhurst one, and as you mention, CK is empty. The publication date for your and Hemingway's copy is 1951, but 1969 for the third one and last entered one. Perhaps LT either takes a random publication date when there's no OPD, or perhaps it takes the PD of the last entered copy?

>173 newcrossbooks: Yeah, none of the books you linked to has OPD filled in the CK so it's taking from somewhere else. If you fill in the right date in Ck it should override any other source.

>149 MarthaJeanne: Maybe I'm being a bit legalistic here, but in my view nationality =/= ethnicity or heritage. It's about what it does / would say on the passport.

For example: my nationality is Dutch, but my ethnicity is mixed, and my heritage is mostly Indonesian

>177 Nicole_VanK: And then there's the issue of the (rather large amount of) people who have dual/multiple nationalities - and the passports to back it up.

>178 PawsforThought: That's one of the reasons why multiple entries are possible, I think

>178 PawsforThought:, >179 Nicole_VanK:
To link up the discussions, we've been talking about the nationality issue over here, too:
https://www.librarything.com/topic/334568

>167 spiphany: AFAIK it's not new and it is intentional, based on what timspalding told me.

BUT, re: https://www.librarything.com/topic/327497, yes! Thank you! I was trying to find this thread yesterday, because I recalled the bug, but couldn't find it for the life of me. So, maybe there's more to investigate here.

"I had this issue earlier today but later on made other changes that do seem to have saved..."

Okay, so you made changes in a non-English CK field and it saved correctly? Do you remember the specific book/field? Are you sure you edited on .com and not the relevant site?

I wonder if you simply have to be logged in on the relevant site, in your browser, but don't necessarily have to edit directly from the site... I'll want to test this further and double check with Tim that there's no bug at play.

>181 kristilabrie: How long has it been that you can only edit other language CKs when logged into those language LTs? Because while it’s been a while I’ve definitely made changes to other languages and be used anything but the mothership site.

>181 kristilabrie: If this limitation on editing CK is intentional, it is new.

As long as I can remember it has been possible to use the language list at the top of CK to edit fields in languages other than the website one is logged into.

I use the German and English versions of LT; I was fixing something recently on some other language version (I'm afraid I didn't make a note of exactly what) and the changes seem to have saved, so the bug/intentional restriction doesn't seem to be consistent.

To be honest, I can't see any benefit to restricting this, and some clear disadvantages. There are plenty of LT users who are not monolingual and are capable of editing CK in multiple languages. It also sometimes happens that obviously wrong information has been entered in another language's CK and if users have to either login on yet another version of LT or ask other users for help correcting the information, that's a huge disincentive to work on CK at all.

In your new presentation of information, there appears to be a major error in the Vous et Nul Autre computations. The last time I checked (recently), I had nearly 1,000 unique titles not shared with any other members, the computation today is 196. I have been collecting vintage public domain e-cookbooks (and a few paper print cookbooks) which no one else has.

In addition, the Page 1 2 3 4 5 6, not the "Shared by Other members" boxed section, doesn't seem to show additional pages of my "only my" books. Whatever is being displayed does not seem to correlate very well with my collection. I notice as I click on the numbers, the number correlates "sort of" with the outlined numbers, though I'm sure the numbers are totally wrong.

However, it is the titles unique to me (Not shared with others) or shared by only 1 or 2 other members which truly interest me. I sincerely believe there is something wrong with the numbers being displayed.

RecipeClippings

>173 newcrossbooks: Where is Original Publication Date being calculated from?

Yes, I need to expose this data. It's being calculated from all the records that go into. Basically, the lowest one gets picked, with some wiggle-room.

If this limitation on editing CK is intentional, it is new.

CK has the ability to make some fields international and some fields not. Most fields are split by language, but some aren't. Non-national ones include:

gender, birthdate, date of death, original publication date, canonical DDC, canonical LCC and some of the legacy library fields.

>186 timspalding: CK has the ability to make some fields international and some fields not.

I got that. There are two concerns here:

1) Whether it is possible (as in the past) to edit those CK fields that are split by language from a different language site (e.g., editing French CK for birthplace or occupation when logged into the English site or whatever).

The bug I linked to in post 167 was a problem with this. Editing across languages seems to be working at the moment, but I had a field yesterday that refused to save my changes until I logged on to the English LT site.

2) Users having difficulty editing non-national fields. I recall there was a discussion of an issue with "gender" not being recognized in statistics if it was edited from one of the non-English sites, which seems less than ideal.

The issue with the original publication date for "Lassy" (posts 155 through 157) also seems to be related to this. I don't know if it's a problem with holdover data in non-English sites from when these fields were split by language, or whether it's the same problem as with editing gender, namely, that changes won't save properly if made on any site other than the English one.

>187 spiphany: I have, a couple of times the last month, changed a birthdate but have needed to go for a hunt to find exactly where I need to change it, because changing it in one place didn't do the trick. Server lag, perhaps?

>184 RecipeClippings:

See >162 newcrossbooks: for more details on this problem.

I think the 196 you are seeing is the number on the first page of the results - it's not totalling up the numbers from all the pages of results.

And the link to the second (or third or fourth) page of results of your Vous et Nul Autre books sends you to the second (or third or fourth) page of results for your You + One other books....

Edit: Seems to have been fixed now

>189 newcrossbooks: Yes, it appears to work now, providing me with the numbers I was expecting

I'm going to look into non-English gender. I think it may be converting it in the "last mile."

It looks like all of my comic books & graphic novels are listed as "Not Set" for fiction/nonfiction. This seems like a strange coincidence for ALL of them to be in there. Video Recordings too but that makes a bit more sense.

Also, I would like to fix the others that are in that category but I cannot for the life of me find where that is in the work record....

All of my comics were "not set" for fiction/nonfiction as well. I did 'em all by hand on the work page.

I discovered the option of editing genre only after struggling with the Not Set books on the chart - so yay, better data through charts!

It is an option as a field in your catalog, under Classifications. There's also Genre (Detailed), which has exactly the same data plus a bar under each entry, with the right end blue. I can't tell what that's supposed to be saying.

Small bug(ish) on the Work Duplicates page - the text at the top says "This page collects all LibraryThing works of which has more than one copy." Of which the user has? of which you have? for which you have? Something other than what's written there.

Is this a bug?

I'm using Chrome on Android. Clicking on any part of the genre graph doesn't bring up the label for that segment, like it does when I click on parts of other graphs. Seems like it should, when many of the labels around the edges get cut off too; all that shows for some of those is "ture", "tion" and "nce" etc.

There seems to be a problem with the height of my book stack. Looks far too high. I have about 5000 books and the stack is supposed to be 2,370 feet. (Or 722.4 m). That would mean each book was about 16cm thick. The reality is that I have roughly 100 m of shelves for the 5000 books so average thickness is about 2cm. (Or am I doing something wrong?)

The graphs are a great new feature. Love them. But noticed that under Genre I have a very large section supposedly about religion. Well I do have some books about religion but I think Philosophy must have been lumped in with Religion. (Commonly done in Libraries but quite wrong in my view). Philosophy is (generally) not religion. I notice in the following chart Philosophy and religion are broken out separately, which gives a much more accurate picture.....just don't see why the genre chart doesn't separate the two.

>198 booktsunami: That probably means when you imported data, it imported incorrectly (or into the wrong fields.)

>198 booktsunami: Add the dimensions as columns to one of the view styles on your catalog and look through your books - the graph shows only what is in your books so fixing it will require fixing your data.

>198 booktsunami: You've got nearly 1600 books where the dimensions have clearly not been entered correctly - the length or height is being reported as the thickness (all seem to be from Amazon data).

You've got seven books with reported dimensions of 20 x 20 x 20 inches.

As >201 AnnieMod: suggests, add dimensions to one of your view styles. Then sort your books by thickness to find the problem books.

When I look at my Books/Cataloging/Classifications chart I notice it shows lots of instances of invalid code for both Dewey Decimal and LC Classification.

The LC Classification code errors, which occur in about 10% of the entries that have the classification, seem mainly to relate to Unclassified items:
UNCLASSIFIED 769 (P), UNCLASSIFIED 2096 (H), UNCLASSIFIEDS-112
or Minimal Level Cataloged items (MLC):
MLCS 96/11866 (H), MLCS 2001/06178 (P)
with a couple of Comic Books:
Comic Book 01806

There are also many entries that simply say:
Newspaper
LAW
UNC
and some green text entries (!?):
DT388.35 .F45, DT376 .E84

Why are the codes for unclassified and minimal level cataloged items not counted as valid codes? They are LC codes that point to unique works.

How can the invalid codes be corrected?

Not sure if others have noted this but I'm spotting it:

When looking at the Ratings/Reviews/Read Dates chart, narrowed to a specific reading year (in this case 2020), everything seems to stack into January and not spread out with the dates, even with the chart showing brackets for the various dates:
https://www.librarything.com/stats/gilroy/share/u953dc87f.ufcc6e992



In the above share picture, you see the lines, though the picture cuts off the labels.
The two dots on the left are all 26 books I have listed as read in 2020. The one dot on the right is where I rated one in 2021.

The actual reading dates over time chart is just as bad:
https://www.librarything.com/stats/gilroy/share/ue519b40a.u47e9ca4e



Clicking into it shows that this should show 12 different columns, not one long extended one...

>196 jjmcgaffey: and >204 gilroy: These should be fixed in a few minutes; I just pushed a change.

>205 knerd.knitter: Much better.

>194 norabelle414: Can you point me to where you did that? I feel like I'm probably looking right past it but I cannot see it.

>207 leahbird: On the right side of work pages, you should see a "GenreThing" box where you can edit/add Genres.

Just stumbled across this bug:

1) select a filter to a specific collection - For me, I was using Location : Dining Room
2) Click into Books, then Measurements
3) go to the Data/No Data Chart
4) Click on any of the options

This takes you to All Collections all data instead of the filtered data that was from on the graph pages.
Also important from this is that when you try to narrow using the collections up top, the grid ignores that drop down, no matter what it's set to.

Something funny seems to be going on with split authors on the Nationality map and list. Phil Barker shows up as "Not set", although the only split I've got any books by has the Nationality field set. OTOH, the seemingly analogous Gunilla Jonsson shows up under Sweden.

>207 leahbird: Yeah, what >208 kristilabrie: said. Opened each book page in a new tab and changed the genre on the right side of the page. I'm still not clear on whether that is how we are supposed to change genres, or from the catalog, but it seems to work for the purposes of my statistics/charts.

I also had to un-genre a whole lot of my comics that were incorrectly classified as "teen"

>209 gilroy: Thank you for finding that; we will look into it.

>211 norabelle414: "supposed to" - both ways work, it's just how much work you want to put in (or which one feels more like too much work!). I made a Genre column in my catalog; if I double-click in a box, I get the list with checkboxes and can select the appropriate ones (and deselect the inappropriate ones). I haven't flagged anything; nothing I've seen so far has be _egregiously_ wrong.

On the Original Publication Dates graph, the "no date" list does not update when original publication dates are added to the works.

I added original publication dates to the two works below on 23 August 2021, and both of them are still being listed as "no date" works in my catalog, three days later.

https://www.librarything.com/work/26444758/
https://www.librarything.com/work/25629987/

>214 shadrach_anki: Yes, it's not clear how the Original Publication date is being calculated. All of my books have a Publication Date but less than 50% have been given an Original Publication Date. And the Original Publication Date that has been given for some books is clearly wrong.

See message >173 newcrossbooks:, replies >175 anglemark:, >176 PawsforThought: and Tim's reply >185 timspalding:

>215 newcrossbooks: Even reading the replies, I'm still not clear.
I went into CK and saved the correct OPD yesterday, and it still appears wrongly in the graphs.

For example, these 2 from 2013-2014 are appearing as OPD 1600-1609!
https://www.librarything.com/catalog/ScarletBea&collection=-1&filter=col...
Smiler's fair
The rough guide to economics

>216 ScarletBea:

I've just corrected a number of books dealing with World War I that had original publication dates in 1600-1699 and 1700-1799. Historians that could see into the future! Also my Original Publication Date with no value has not changed in a week even though I have made many corrections.

Another nice idea that has been badly implemented.

In the cataloging section for identifiers it has a section - invalid identifiers. When you click on that part of the graph you get your books that are classed this way. According to the output my ISBN's are wrong for these books but a random check indicates I put in the ISBN number as listed in the volume correctly.

As an example the output states my ISBN for Niven's Inferno is wrong but the number in the book matches what I entered - 855232714

>218 alco261: That's a nine digit number. ISBNs are 10 or 13 digits, so that's an invalid number. It looks very much like you accidentally left off a leading zero.

>218 alco261:, >219 Foretopman: In all fairness, I suppose it's possible the publisher accidentally left of a leading zero, but a publisher's mistake still doesn't make it a valid ISBN.

>218 alco261: >220 Foretopman: That's what appears to have happened. Allan Wingate left off the leading zero.

https://isbnsearch.org/isbn/0855232714

When ISBNs were new, a lot of books didn't bother with the leading 0 (or possibly it got formatted away? fewer computers to blame then, though). 0 if it's printed in the US, 1 in the UK, I don't know the others. I have a lot of books (especially SF paperbacks) with 9-digit ISBNs - I don't even think about it anymore, just plug in the first digit.

There were SSBNs before ISBNs, and they were 9 digit numbers without that first 1 or 0.

I seem to have a lot of invalid ISBNs, but I don't really feel like trying to fix them.

That would be why. I knew that, once, but I'd forgotten. Yeah, S Book Numbers (what's the S? Ah, Standard) and then added the country to make it International S Book Numbers.

>220 Foretopman: Thanks - that has to be it - there's no 0 printed in front of the number sequence in the book.

Is the button for measurements not sticking? Or is it tied to the device you're using? Because I changed to metric as soon as it became possible but now that I checked it was back to imperial. I'm on a different device now, so could be that but you shouldn't have to click every time you're on a different device.

Also, the "Totals & Averages" is still only showing imperial even if you've chosen metric.

Also, there is still a problem with Book covers. My charts for "Where Are Your Covers From?" shows 36 books, even though it's set to all books in all collections. I have more than 36 books - why is this happening?

>226 PawsforThought: I want to say that the book covers graph shows only those books you have not yet personally interacted with. So you have 36 books where the system is determining the cover. You haven't.

At least that's what happens when I look at my graph.

>227 gilroy: No, that's not it. I've looked through the list of books from that chart and it's either books I've chosen a cover that was already uploaded or a cover I've uploaded myself. It's not auto-chosen.

>228 PawsforThought: Mine are exactly the way gilroy says. There's only 8, all automatically selected ones. Either way, seems buggy

>229 SandraArdnas: No matter what it is, a chart called "Where Are Your Covers From?" where the centre point is call "All" should include *all* books in your library. Otherwise, what's the point?
As it is now, my chart looks like this: centre (full circle): All / first circle (full): member uploaded cover / second circle (full): best guess for ISBN (full) / third circle (near full, save one book): high quality upload (the one book is low quality)

I have no chart piece for chosen by me, even though a large part of my covers are chosen by be (I'd guess the majority). A large part of those are even uploaded by me, but there's nothing showing that in the chart either. Something is very buggy, indeed.

>226 PawsforThought: The Metric/Imperial flag is cookie-based, so it will stick as long as you're on the same device/browser.

ETA: Also, the Totals & Averages table has always only shown Imperial.

>231 knerd.knitter: The Metric/Imperial flag is cookie-based, so it will stick as long as you're on the same device/browser.

I figured, but why can it not be user-based instead? I'm unlikely to want to revert back to imperial just because I'm on another device.

ETA: Also, the Totals & Averages table has always only shown Imperial.


Yes, I know. And I don't think it should. It's useless to people who don't use imperial.

>232 PawsforThought: My point is that both of those are working as intended and are not considered bugs.

>233 knerd.knitter: It's a problem, bug or not.

>233 knerd.knitter: I think you're looking at the wrong feature. If the imperial/metric toggle was intended to apply to the whole site, and it seems logical to assume that, then having it not apply to one particular table is a bug.

>226 PawsforThought:, >227 gilroy:, >228 PawsforThought:, >229 SandraArdnas:, >230 PawsforThought:

The covers problem was reported as a bug two weeks ago:
https://www.librarything.com/topic/334497

It's not fixed yet....

>235 jjwilson61: That toggle on the measurements page used to only apply to the pages/weight/height/thickness charts; now it also applies to the how tall is your book stack because that was a fairly simple change.

>237 knerd.knitter: That's great to finally see them in measurements that I understand.
Would it be possible to also have this affect the height and weight totals on top of that page?
It would also be nice to have the volume in liters or cubic metres instead of bathtubs.

Thanks!

>238 igorken: There is volume in cubic meters, bathtubs are just fun.

However, choosing metric does not affect height and volume at the very top row of 'measurements' page. Is this a bug? Pages show properly in kilometers. Weight is curiously in stones, which I wouldn't expect to be default for either metric or imperial.

>238 igorken: >239 SandraArdnas: The values at the top of the page are randomly pulled from the values displayed under Facts at the bottom of the page. All the data is there.

ETA: The values other than the height one. That is the value from the How Tall is Your Book Stack chart.

>240 knerd.knitter: OK. I'd rather see the ones on top in a scale that I select as I find the quirky things fit better in the table at the bottom, but it's a very minor thing. It's all there at the bottom indeed.

>240 knerd.knitter: Ah, OK thanks. Haven't noticed it's random values

It would be nice to be able to see all my authors by gender. The "full list" seems to only be about 500.

>243 MarthaJeanne: Simple fix; should be there in a few minutes.

>244 knerd.knitter: Oh, Wow! Yes! Thank you!