Round the hill

A month ago the ITO OSM analysis was showing my local area as having reached 74% coverage. It ranked about half way up the list, with roughly equal numbers of other areas showing better and worse levels of coverage.

Checking out some of the inconsistencies has given me plenty of excuses for pleasant outings over the last month. Several bank holidays later we have just passed 90% coverage, and risen to rank three-quarters of the way up the table. About a quarter of local areas still have better coverage, and of course, several of them are also improving, so the bar is continually rising.

That still leaves a gap of 10% to close, and some of the remaining inconsistencies are a bit of a puzzle. This is one. With that sign, the OSM name tag on the road was previously set as "Knowl Hill Common (Round the Hill)". I'm not sure, but I think it might have been my contribution in the first place. It certainly seems a reasonable interpretation to me.

However, on OS Streetview this road is named "Knowl Hill Common". A cursory search of the internet suggests that local addresses also tend to prefer "Knowl Hill Common". On the ITO analysis, half of the road was highlighted as inconsistent, and the other half wasn't.

After mulling this over (note to self: must get a life), I decided to change the name tag on OSM to "Knowl Hill Common", with an alt_name tag of "Round the Hill".

Of course the beauty of all this is that anyone who feels strongly, or has better local knowledge. can always change it back, or change it to something different. The location is here.

At the same time, and again thanks to ITO, I've uncovered a couple more named roads that I'd previously missed, despite having ridden past them regularly for the last three years. I can't believe there are so many of these bits and pieces that I hadn't noticed before, but in my defence, both of these look a bit like private driveways from the road. Closer scrutiny on the ground shows that they do have real names, and both are now fixed.

And finally, to my relief, I've managed to fix a mess that I made of one of islands in the Thames. I screwed up one of the riverbank multi-polygons last night. I realised shortly after that this had drained a short section of the river upstream from Windsor. It wasn't a difficult mistake to fix, but it was a bit of an uncomfortable wait before the solution was rendered again. All seems to be well now.

Onward and upward.

Somebody tried to steal my old bike last week

We only have room in the shed for one bike, and the space is taken by the touring bike that I got last year. My older hybrid used to live in the shed, but now it hangs on a rack round the side of the house. It's outside, but protected from the worst of the weather, and secured by a padlock and wire cable.

On Friday morning we discovered that overnight somebody had removed the cover, and lifted the bike off the rack. However, they must have been deterred by the lock, and the bike was still there, lying on the ground.

The bike wasn't visible from the street, and we are tucked away on a side road, out of the way of passing traffic. So it's a mystery what anyone was doing at the back of the house in the first place. Presumably they had come over the garden fence looking for a short-cut, seen the bike, and thought "I'll have that", before realising it was going to be more trouble than it was worth.

We keep the second bike for visitors, and I use it occasionally - such as when my main bike was in the shop for repairs. Most of the time it just hangs there gathering cobwebs. That seems a shame, so rather than just putting it straight back, I've cleaned it up today, checked it over, pumped up the tyres and taken it for a ride. It's the first time it's been out in months. After more than a year using a new bike the differences in the ride were obvious straight away, but it made a change, and I soon adapted. I'm probably kidding myself, but when I clean a bike I can usually convince myself that it is running particularly nicely.

I went a few miles out to Knowl Hill, looking for a few things that I knew need fixing on the OSM map, but more of that shortly. In the meantime I might not have made best use of both bikes over the last year, but I'm glad that I still have the option.

Wedding present

The spectacle has been impressive, but the coverage is becoming too much to bear.

Nevertheless, who could do anything other than wish the couple well? So I spent a bit of this morning creating a wedding present for them.

This maps the loyalty of the nation, according to the proportion of street names with royal connections. On the chart, yellow presents a high proportion of royal street names, and red represents a low proportion.

The most popular examples of street names with royal connections include Victoria Road, King Street, Queen Street, and various other combinations of Victoria (...Street, ...Park, ...Avenue). Princess Road and Princess Street come next, ranking seventh and eighth by total length. Together these eight names account for about half of street names with royal connections. King Edward Road ranks ninth, Queen Elizabeth Way ranks tenth, and Prince of Wales Road ranks eleventh by total length.

Altogether royal street names account for about 0.05% of the total.
The highest proportions are in the City of London and Denbighshire. The lowest proportions are in Harrow and Peterborough.

The street name data came from the OSM database, and the boundaries from Ordnance Survey. Postgresql, Postgis, Open Office and Quantum GIS did the heavy lifting.

Sustrans FAQ

Sustrans are great, but their online mapping isn't, and this FAQ response doesn't make sense to me....

Why don't Sustrans use Google Maps or OpenStreetMap?

A Google or OpenStreetMap API (a map that can sit within another web page) re-loads the data attached to it each time you move or zoom the map. Our dataset is over 80Mb so this would run very slowly.

We use Ordnance Survey 1:10 000 and 1:25 000 scale maps as our backgrounds for several reasons:

They are based on the same map product and so will match up with each other at each zoom level. This means that routes will appear correctly against the different backgrounds. For example, a route plotted down one side of a carriageway against the 1:25 000 will appear on the same side of the carriageway when viewed against the 1:10 000 background.

Google mapping is not consistently detailed enough in all areas of the country for our needs. Also areas of shadow created, for example by tree-lined lanes, can obscure important information for cyclists and walkers. We believe that Ordnance Survey backgrounds give much better detail in both rural and urban areas.

From http://www.sustrans.org.uk/about-sustrans/faqs

Which councils chose the most boring street names?

The widest variety of different street names is to be found in central London. Manchester, Scottish cities, and some English ports also show quite a lot of variety in their choice of street names. The least interesting mix of names tends to be found in rural counties to the east of England. East and North Yorkshire, Norfolk, and Lincolnshire all have a high proportion of common names, and use little variety in naming the rest of their network.

The bottom line is that the most interesting names are in Islington, and Kensington & Chelsea. Norfolk must try harder.

Allowing a bit of slack for spelling variations and tagging errors, the OSM database holds well over 250,000 different street names across the whole of the UK network. But some are more common than others. Only 25 of the most popular street names account for almost 5% of the named network.

"Station Road" and "High Street" are the most popular names. These represent respectively just over, and just under 0.5% of the total. "Church Road", "Church Lane", "Main Street", "Mill Lane", and "London Road" are also among the 25 most conventional names. However, the pattern is not consistent. In parts of London and Scotland the most common names account for less than 1% of the network; while in other parts of the country the 25 most common names account for almost 10% of the named road network.

Outside the most common names, the bigger authorities tend to use a wider variety of names - as one would expect because they operate bigger road networks. The big counties, like Kent, Essex, Lancashire and Hampshire each uses more than 1,000 distinct street names. On the other hand island councils such as Orkney, Shetland, Western Idles, and Anglesey; and smaller authorities on the mainland use a much smaller number of different names. Adjusting for the size of each road network, central London boroughs, and some cities, such as Stoke, Manchester, Bristol, and Derby each use a wide variety of different street names in proportion to their size, while shire counties generally show less variety.

The local authorities with the highest proportion of roads using common street names are:

• East Riding of Yorkshire
• Norfolk
• Moray
• Bedford
• Lincolnshire

The local authorities with the lowest proportion of roads using common street names are:

• Lewisham
• Western Isles
• Islington
• City of Westminster
• Kensington and Chelsea

The local authorities with the most diverse mix of different street names are:

• City of London
• Tower Hamlets
• Stoke-on-Trent
• Islington
• Kensington and Chelsea

The local authorities with the least diverse mix of different street names are:

• Norfolk
• North Yorkshire
• Lincolnshire
• Western Isles
• Orkney Islands

And as exercises in unnecessary and meaningless statistics go, I reckon all that will take some beating.

More dreadful weather for armchair mapping

The weather is glorious again today, so I set out early and took the scenic route to Windsor. I pootled around Dedworth for a while writing down some of the street names that are missing or incorrect on OSM, so I could add them when I got home. Then I rode back through Bray, with a stop at Boulter's lock where I treated myself to an ice cream and watched the boats for a while.

Dedworth is a suburb of Windsor, which happens to lie between some of my regular rides. It had quite a few missing street names, and it's a reasonable distance for a ride of a couple of hours. Those are the reasons I picked it out, not because I had a burning desire to visit Dedworth. It seems to be made up of a series of developments that mostly date from the 1930's to the 1970's. For example, here's a Pathé news clip about some self-build activity here in the 1950's.



While I was updating the OSM data I started to get curious about some of the street names, which are a bit unusual. It turns out that Dedworth has Saxon origins, and predates Windsor. Although the housing is relatively modern, they have used some names with historical associations. Examples I logged today include "Filmer Road" (Filmer was one of the Windsor Martyrs, burned at the stake for heresy in 1543), "Frymley View" (Frymley was a 16th century mayor of Windsor) and "Surly Hall Walk" (Surly Hall was a local inn, used by pupils at Eton). I also added names to some cul-de-sacs off Gallys Road, which is named after the landlord of the Garter Inn, who appears in Shakespeare's Merry Wives of Windsor.

The history of Dedworth: not as boring as I thought.

Traffic delays

I've ridden past this sign a few times.

Spotting it out of the corner of my eye, I thought it read:

Traffic delays
Likely road rage in this area
Good Friday
9am - Noon


So I stopped to take a picture, and was disappointed to realise my mistake.

"Another unnecessary report showing meaningless statistics"

Around here we can claim the highest proportion in England of children who are taken to school in cars. 46% of children usually travel to school by car, compared to a national average of 28%.

When these figures were published in January the local paper asked whether this meant the town's children were lazy and described them as "cosseted". The council, on the other hand, described the figures as "another unnecessary report showing meaningless statistics" (I think we can assume that they found it embarrassing).

I don't normally cycle at the peak times for travel to and from school. But whenever I cross a school route at other times my impression is that provision of traffic-free cycle paths isn't too bad. I had assumed that they just weren't being used very much.

However, looking at the figures more carefully, it's not the proportion of pupils who cycle that should be of most concern.  Around here it's the proportion of pupils that walk or use public transport that are well below the national averages. The proportion who cycle to school isn't as bad as I had expected. It may only be 4% of all pupils, and that leaves plenty of scope for improvement. But to keep things in perspective, 4% cycling to school around here compares to only 2% nationally, and 3% across the South-East.

It's ironic, of course, that the main reason that parents accompany their children to school is  fear of traffic, and that cars taking children to school represent more than 20% of that traffic at peak times in urban areas. And those are national figures for all urban areas. I don't know what the local figures are, but an awful lot of the traffic that is seen as posing a risk to pupils on their way to and from school must be on the road because it is taking part in the school run.

The figures are on the DfT site. Here for the National Travel Survey and here for the School Travel Statistics. They relate to 2009, but I suspect these might be the last government figures that we see on this subject for a while.

Frightening the elderly

View Larger Map

I rode down this shared-use path this morning. It leads to a pedestrian crossing, where I rejoin the main road, at a short section of cycle lane. This is a bit of a tricky junction for both cyclists and drivers, and in the past I've thought of this shared path as a handy way to keep out of the way of traffic. Presumably that's why it is there.

Normally there's nothing very remarkable about any of this, but this morning, as I approached the pedestrian crossing there was a rather frail elderly lady approaching it from the opposite direction. Seeing me riding (slowly)  towards her she looked absolutely terrified, and froze on the spot until I was back on the road again.

I should probably mention that I was moving at about walking pace, and I was some distance away from her.

She was using a stick, and she didn't look altogether steady, so it is perfectly natural that she would be cautious. But the frightened expression on her face went far beyond what I would expect from somebody who is feeling a bit vulnerable. I stopped, tried to make it clear that I wasn't heading in her direction, and moved off slowly while she remained frozen. Perhaps I should have called out something friendly, but that could have made things worse.

I have no idea, of course, what was worrying her. I don't think I look any more scary than any other (late) middle-aged man wearing a cycling helmet. Maybe she has had a bad experience in the past. This junction is at the bottom of a hill, and it's perfectly possible that somebody has ridden too quickly around the corner and given her a fright. It may just be that she has seen some of the recent press coverage about dangerous cyclists.

Whatever the reason, she has given me pause for thought.

I don't often use shared footpaths. If there are any pedestrians around it's normally easier all round if I keep out of their way. On the other hand, there are places where separating cyclists and cars can also make life easier for both. In places like that I will use a shared path, carefully.

However, it's not difficult to choose between scaring the elderly on a shared footpath, or forcing cars to travel briefly at my speed on the road.

Hawthorn hedge

Somebody has made a lovely job of relaying this hedge near Shottesbrooke. There is a limited amount of blossom starting to appear. I don't know whether I should expect an abundance this year, but already it's a treat to see the foliage flourishing each time I ride past.

Physics of the riderless bike

This is a return to a topic that I've touched on before, and I'm not sure that I'm any wiser this time, but it's a nice little film.

Thanks to Ben Goldacre for the link.

Talatat

"Talatat are stone blocks of standardized size (ca. 27 by 27 by 54 cm, corresponding to ½ by ½ by 1 ancient Egyptian cubits) used during the reign of Akhenaton in the building of the Aton temples at Karnak and Akhetaten.

The standardized size and their small weight made construction more efficient. Their use may have begun in the second year of Akhenton's reign. After the Amarna Period talatat construction was abandoned, apparently not having withstood the test of time."

I came across that by accident, after mistyping "Tlatet" in a search. It comes from http://en.wikipedia.org/wiki/Talatat. But it chimes with something else that I've been mulling over recently.

My programming career started in the days when tools were much less sophisticated, and much less reliable than they are these days. The move from "spaghetti code" to "structured" programming was under way. We were being encouraged to use procedures more, and avoid "goto". But while objects, declarative programming and the like may have been around in academic circles, they hadn't reached the coal face. The big and clever tools weren't as big and clever as they thought they were. It was always helpful to stay close to the underlying components so that when things went awry you could usually figure out why.

I earned my crust programming for a few years, then moved onto other things. By that time things were starting to change, and I've learned about subsequent developments out of interest, rather than for professional reasons.

The point of all this is that since my formative years of programming I've always been more comfortable when I can see how the cogs turn. I'm hugely impressed at the level of sophistication that open software is achieving, and I enjoy learning how to take advantage of it. But often I feel as though I'm dropping things into a "black box" then waiting for the results to appear. When things don't go the way that I expect, I'm inclined to open the box up, take it apart and examine the workings. Normally it turns out that my instinctive approach isn't practical. I know I could get hold of the underlying code, and that's a big advantage over proprietary software. But in practice there's a huge drop down from today's high level of sophistication, to the mass of minute detail that undepins it.

I'm suppose I'm hankering after being able to muck around with building blocks that are smaller than the major software tools, but bigger than the underlying code. We might call them "Talatat". But it seems that they didn't stand the test of time either.

Windermere refilled

Having tidied up the data, tweaked the rendering, and speeded up the whole process I finally got round to regenerating my 1/2" maps, only to discover that Lake Windermere had disappeared. Refilling it has taken a while.

Generating the foreshore from Ordnance Survey VectorMap District was a bit fiddly, but a lot more straightforward, and I think the result is worthwhile. Thanks, Chris for pointing me in the right direction.

There are two more examples here and here.

Cycle infrastructure tagging

My first attempt at using Mapnik to produce a version of the old 1/2" maps was a bit of a bodge. Learning what was possible involved a lot of trial and nearly as much error. In the end I created something that was vaguely like what I'd intended. Now I'm in the process of rebuilding it all in a more organised way.

The rebuild is going reasonably well. I've got better source data, and discovered and fixed some errors that crept in last time round. I've achieved some worthwhile performance improvements and brought the processing time down from a few days to a lot of hours. At the same time I'm tweaking how the result looks.

As part of the whole process I'm trying to find more effective ways of capturing different tagging variations. For example, to find cycling infrastructure I search for various combinations of "highway=cycleway", "cycleway=something", "bicycle=yes", "ncn, rcn, lcn, ncn_ref, rcn_ref, or lcn_ref=something" and "route=bicycle". I currently find about 44,000km of UK cycling infrastructure with those combinations. By way of comparison, Sustrans say that their national and regional network is 12,600 miles (i.e. just over 20,000km) in length. The rest of what I'm finding is made up of local networks, and cycling stuff that doesn't form part of a network.

I end up plotting this at 1/2" to the mile - so there isn't scope for a lot of complication in the way I plot the different options. At the moment I'm characterising all of the different types of cycle path as either "on-road" or "off-road". It turns out that I classify 53% as being on-road. The way I've set things up, that effectively includes primary, secondary, tertiary, unclassified, residential, and other roads that are marked with a cycle-related tag of some kind. I'm classifying the other 47% as being off-road. Of the network that I capture about 3% fails to give any indication of the type of road or path, and I've assumed all of those are off-road.

All of these figures are UK only (more or less - the bounding box overlaps some other countries slightly), and I pulled the data at the beginning of April.

Looking first at the on-road element of the cycle network, only 1% is on roads that mappers haven't classified yet (i.e. highway=road). 8% is on a primary road (plus another 3% on a trunk road). 12% is on a secondary road. The way that lesser roads are classified can be a bit arbitrary, but according to OSM mappers, almost half (46%) of the UK on-road cycle network is on an unclassified road, 20% on a tertiary road, and 8% on a residential road. That leaves 2% on service roads, and a very little bit on different types of link-road.

Elsewhere, six tags account for 93% of the off-road network. The most popular is "cycleway" (almost half of the off-road network). "Bridleway" accounts for 16%, "track" for 11% and "footway" for 10%. The only other common tags are "path" (4%) and "byway" (3%). Among the less common tags we find "pedestrian" and "unsurfaced". "Steps" occurs a couple of hundred times, but doesn't account for much of the network length. More rare tags include "construction", "proposed", "living_street", "crossing", and some combinations of the above.

For obvious reasons I can't measure how much of the cycling infrastructure I'm missing. There must be some, but I've not spotted any obvious gaps with my improved process. Trying to measure the "unknown unknowns" is still on my list of things-to-do when I work out how. Of the infrastructure that I've managed to find, something like 96% is tagged in a way that I can understand. The other 4% are the "known unknowns" - I know they are cycling infrastructure but I'm not clever enough to do much else with them.

Leaving the unknowns aside, at first I found that the variety of information about the rest was a bit daunting. There was a stage where I would spot an apparent gap in the network, and discover that it was tagged with a variation that I hadn't thought to include. I started leaning towards the school of thought that somebody needs to get a grip on standardising the tagging scheme, so that we don't clutter things up with too many variations.

Now that I've got myself a bit more organised, I'm much more relaxed about all that. There aren't too many variations in general use. The different tags that are used widely reflect different characteristics of the infrastructure design ("Highway=cycleway", or "Cycleway=something"), the way it can be used ("Bicycle=yes") or the network that connects the bits together ("NCN", "RCN", "LCN"). I'm much more inclined now to the view that the benefit of capturing that rich variety with flexible tagging is far more important than the inconvenience of having to handle some of the more obscure variations.

Dreadful weather for armchair mapping

"Completeness" is a bit of a moving target when it comes to Open Street Map, but I don't suppose that I'm the only person who wants to see some sort of completeness in the content that matters to me. At the moment what that means in practice is that I would like to see an end to the coloured splodges that show up on the ITO analysis of my local area.

Like others, I've been steadily nibbling away at the missing roads, and the analysis is slowly turning monochrome, but unlike the most organised parts of the country, around here there is still plenty to do.

Because I'm keen to clear up the discrepancies there's quite a strong temptation to take the easy way out, and fix things without getting out there and taking a look. In some cases that seems to be the sensible approach. There doesn't seem much point in taking a round trip of a few miles to check that it's OK to change a name from "Abbreviation Rd" to "Abbreviation Road". Or to make sure that the right name is "Spelling Avenue", not "Sppelling Avenue".

On the other hand, without taking a look it would be easy to miss some of the genuine inconsistencies between the OS data, and what can be found on the ground. So I try to resist the temptation to do too much from the armchair. I try to keep a sensible balance between fixing things the quick and easy way when the reality of the situation is pretty obvious; or going to take a look myself when there is some doubt.

This morning was a lovely day. As a result the balance swung strongly in favour of getting the bike out and riding out to take a look around. The area I decided to cover was some suburban housing to the west of Windsor. It's not the most interesting destination, but there were quite a lot of coloured splodges in a small area that lies just off one of the routes that I regularly take purely for nice ride. On a loop of about 20 miles I managed to fix about a dozen road names, as well as enjoying a very pleasant ride out there and back.

Most (if not all) of the fixes could probably have been done easily without a visit - but what a waste of a fine day that would have been.

I'm still not sure what to do about discrepancies that are difficult to check - even with a visit. A number of those are either a country lane for which the OS has a name, but there doesn't seem to be any name displayed in situ. Or they occur where a road name changes, but it isn't clear exactly where.

I could assume that the OS knows what they are doing, and that the original tagging was wrong - and change the name accordingly. Or I could assume that the original tagging was based on local knowledge, and that the OS has got it wrong, and tag to suppress the error report. I suppose somebody will have a definitive answer, and I could try to research it - but that seems a bit over the top for short sections of insignificant roads. At the moment I am leaving cases like these for someone else to sort out.

So, someone else, whoever you are, the weather outside is lovely, and it's the perfect time to pull your finger out, and fix some of the stuff that I'm leaving. Until you do 100% "completeness" is going to be unachievable.

Cycle hire virgin no longer

It's been a good day today. Not least because, at long last, and months after my key arrived, I've managed my first ride on one of the London hire bikes.

The weather was warm, and I needed to get from Holborn to Paddington. As someone who isn't used to riding in the middle of London, I bottled out of using the docking station on High Holborn because of the traffic. I had walked up to the British museum before I felt that I should stop being such a wimp.
I fumbled around a bit, but after a couple of false starts, the undocking process was straightforward enough.

I'm not exactly trailblazing here, so the ride itself was much as expected. My route finding to Paddington was a bit hit and miss, but by the time I arrived I felt I was coping OK with the traffic. It helped that nearly everything on the road (apart from a few fast cyclists) was moving at a slow pace, and my rather sedate progress didn't feel as out of place as it would around here.

I didn't remember to keep careful track of the time, but I think I got the bike docked within my free half hour, despite having to search a little bit for the Paddington docking station.

So if there is anyone else left who hasn't got round to trying them yet, the traffic isn't as scary as expected, they are not at all bad to ride, and the system is easy to use. They have my vote as a practical and fun way of covering a few miles across the middle of London. I still prefer not to think of them as B*r*s bikes though.

Holes

When I was trying to produce contours for my homage to the old 1/2" maps I discovered that there are "holes" in the SRTM altitude data. The software I was using doesn't seem to handle them very well, and they produced some strange effects, but those were quite localised. As far as I can see, all the problems were well away from the main areas of the UK that I was working with. So I added this to the list of things that would be interesting to explore later, and I ignored them.

Then today I came across this on the Ordnance Survey blog.

It's about producing a list of holes in Britain, as an addition to the existing lists of mountain peaks such as the Munros. My immediate reaction was that this must be related to the same problem. It isn't of course, but the video was worth a view anyway.

"It's impossible to feel like a grown-up when you're on a bicycle"

There's a wonderful article by P. J. O'rourke on the Wall Street Journal. Thanks to @carltonreid for the pointer.

Not everyone who has commented seems to realise that the author is a satirist. In other words, his job is to use irony and wit to expose human folly.

Like this:

"Bicycles are the perfect way to go nowhere while carrying nothing"
"The reason it took mankind 5,000 years to get the idea for the bicycle is that it was a bad idea"
"You can walk up three flights of stairs carrying one end of a sofa. Try that on a bicycle"
"You can't decrease traffic congestion by putting things in the way of traffic"

Those are jokes, not statements of his core beliefs.

I don't know whether P. J. O'Rourke cycles or not. I suspect he might. This isn't his first foray into similar territory. And when he says "It's impossible to feel like a grown-up when you're on a bicycle" he get pretty close to part of the appeal.