Spider maps update - still a mess

Open Data
1

The spider maps section of the TfL Data Bucket saw an update on 13th November. I was hoping that this would solve the mismatch with the website but it has not. The following uses the Excel file that I have placed at

http://www.timetablegraveyard.co.uk/Spiders_update.xlsx

The Data Bucket lists 1163 spider maps. Of these, 708 have a creation date of November 13th and 455 show May 28th.

The website references 545 different spider maps (after deduping). All of these appear in the Data Bucket with the creation date of November 13th,

Thus the Data Bucket contains 618 files not visible on the website, consisting of the 455 with the earlier creation date plus 163 with the later creation date.

Past experience suggests that there should be very roughly 1000 live spider maps. It is therefore no surprise to find two different files in the Data Bucket for some locations, sometimes (but not always) with the same creation date.

It should be noted that there are files with both creation dates that are missing from the website, e g barking-longbridge-road-a4-300917.pdf with the earlier date and southgate-a4.pdf with the later one.

The different form of the urls is to be noted. The Data Bucket urls are of the form

http://bus.data.tfl.gov.uk/bus-route-maps/abbs-cross-a4.pdf

while the website equivalent is

https://tfl.gov.uk/cdn/static/cms/documents/bus-route-maps/abbs-cross-a4.pdf.

So it would seem that the Data Bucket is listing files from somewhere other than the live location. That would explain inconsistencies but not the denuded state of the website.

Slightly off the point but the total lack of naming discipline for the maps stands out. Sometimes the date is included (not always in the same format). Sometimes “a4” is included. Sometimes an extra “and” appears, the use of hyphens is inconsistent and parts of the name may or may not be abbreviated.

There is an argument for including a date in order to distinguish different editions but the other differences don’t look like good practice.

All this means that third parties linking to a particular map may find their link ceases to work. It should not be assumed that everybody uses the same approach to get at open data!

More trivially (while i am at it), the website shows a Brentford map under Brent and two Bromley (by-Bow) maps in Bromley borough. Thankfully they are also shown under Hounslow and Tower Hamlets. I also note that WPR apparently play at Lotus Road…

2

The bucket list of spider maps was updated again on 21st November. It now shows 630 spider maps rather than 1163. Much (possibly all) of the duplication has gone. So too have a lot of maps which should almost certainly exist - Albany Park, Battersea Bridge…

631 of course still exceeds the 545 figure from the website itself so it would appear it has not just been brought in line. I have not done a full check of the website but I can confirm that Brent Park has a map in the bucket list but is not present on the website.

The fundamental problem that the bucket and website point to totally diffferent locations is still to be addressed, months after I first pointed it out. Nobody has said they are synonyms for each other (and if they were there ought not be any discrepancies) so I presume they are not.

Michael

3

@mjcarchive

Is there anything that us end-users can do to help TfL out here?

4

Brian

Thanks for replying. I was starting to think nobody anywhere was reading my posts on this.

As nobody from TfL says anything about the cause of the problem or actions they have taken, I have no idea whether the last update was the first stage of a reaction to my analysis last week. Some feedback from TfL would be useful.

I would hazard a guess that at least 300 locations which have had spider maps in the past are no longer represented on the website or (now) in the Data Bucket. I suppose the possibilities are:

  • TfL no longer wish to present spider maps for these location;
  • The files exist but they are in some (unknown?) third location;
  • The files have been accidentally deleted altogether somehow

Possibly end users could supply the relevant files for the last two of these cases but they might not be up-to-date. I could have a stab at producing a list of (probably) missing locations. But I see little point in devoting more effort to it without feedback on what the problem is and what might be needed to sort it out.

When it comes to the disconnect between the locations linked to from the Data Bucket and the website, there is nothing the end user can do other than draw attention to it. Given that the role of the Data Bucket AFAIK is to list the files that are available via the website, I find it hard to see how it can have been allowed to develop in the first place.

Michael

5

@mjcarchive I read everything on here, but using my RSS Reader, so I only comment when I can add something useful.

I can’t help remembering that the Mayor was going to change the spider maps to use the same colours on all maps [1][2] but perhaps that is before he starved TfL of money to keep the fares from going up?

My personal guess is that they spider maps count as “artwork”, rather than “data” as they require the intervention of a graphic designer for each one and there just is the cash in TfL to get these sorted out - with out without consistent colours - even though you do see them stuck to bus-stops.

[1] London's red buses to be colour-coded to match route maps | London Evening Standard | Evening Standard
[2] https://www.citymetric.com/transport/four-thoughts-tfl-s-plans-colour-code-its-bus-routes-3001

6

Brian

Thanks. It was the lack of any response (and interest?) from TfL that I was getting at.

The colour coding was, perhaps still is, a trial in two areas and I don’t recall seeing the outcome of the trial reported let alone any hints of Londonwide extension. I did see one colour-coded bus for the 247 … parked in the Forest of Dean. Sold but not repainted!

The form of spider map that accompanied this trial was pretty awful (see http://bus.data.tfl.gov.uk/bus-route-maps/barking-a4.pdf which is only available through the Data Bucket), losing the last semblance of geography in favour of a belief in long straight lines. But that is getting off the issue…

… which is the chronic inability to maintain and update. Staff cuts perhaps? Having two distinct and inconsistent streams cannot help.

Michael

7

Bump…

Still no improvement on the website, let alone the Data Bucket.

1 Like
8

The Data Bucket spider maps were updated, or at least overwritten, on 11th December and there are now still fewer, at 594, which I reckon is at least 300 light.

Needless to say, the set in the Bucket does not match that available via the TfL website proper.

I’ve been banging on about this for months but nothing happens. I would have been embarrassed if this had been the position for a day, let alone months, for something for which I had responsibility. Sorry to put it this way but does anyone at TfL concerned with spider maps actually care any more?

Michael

1 Like
9

Following the 20th January update of spider maps in the Data Bucket, I have looked in more detail at the comparison with the map pages on the TfL site.

As before, there are locations (45 of them) present in the Data Bucket but not on the Map Pages. None the other way round.

I found 79 mismatches, where the files for a location differed. In 71 cases, the Data Bucket file was older. In 8 cases it appeared to be newer but I think the files are the same (apart from a date change).

To my surprise, of the 412 files which matched, 79 had different file names on the two sources.

The Data Bucket contains 53 duplicate locations. Either there are two identical (but differently named) files or a redundant older version is present.

I also looked at locations not present in either course. I found over 400 such locations for which spider maps have existed in the past, including ALL spider maps in the Putney, Streatham, Wandsworth and Willesden areas.

I can provide a full list of these locations if it would be useful; also examples of the various form of inconsistency discussed above.

Michael

10

I’m wondering if what going to happen to the spider maps is they are going to make until an automatic “spider-y” map can be created. Like the ones outside Liverpool Street at the moment … must be getting towards “codable”, rather whatever it is that the big spider maps show.

11

It had occurred to me that they might be allowing them to die slowly but some of the maps missing from the map pages (and the data bucket in most cases) are recent and unlikely to be incorrect, so why are they no longer available?

The next to useless August 2018 maps for the Hayes area were (I hope) just an experiment. The idea seemed to be to mimic the approach used in Tube carriages but it simply does not work with bus services which criss-cross each other repeatedly. The more usual spider maps are not geographical but you do get some sense of direction from them.

Although the details of the issues may be different, there seems to me to be a parallel with the Working Timetables issue; if there is clear guidance about how file transitions are to be dealt with, they don’t seem to be followed very well. In the spider maps case it seems we have two totally independent streams in which files can be named differently in each stream. I have commented before on the absence of any consistency in the file naming, which does not help.

Michael