Address Twins: One man’s search for the shortest distance between two homes with the same address

Another popular street name. Image: Getty.

If you lived in a house with the address ’5 Station Road’, it probably wouldn’t surprise you if you learned that there was another house somewhere in the country which also had the address ’5 Station Road’. There are, after all, over 2000 streets named ’Station Road’ in the UK.

But the fact that you sometimes get streets quite close together with the same name can be confusing. In London, there’s a sign on the platform at Abbey Road DLR station telling tourists, via the medium of appalling puns, that if they’re looking for the Abbey Road made famous by the Beatles, they are, in fact, in the wrong place. The Beatles’ Abbey Road is 12.5km away from the Abbey Road DLR station.

I found myself wondering what the shortest distance was between two Address Twins – houses with the same number and the same street-name.

Unfortunately, I can’t find a publicly-available list of every house in the UK – Royal Mail have house address data, but you have to pay a minimum of £399 to licence it. But I do have an Ordnance Survey list of every street in the UK, so my plan was to find every pair of streets with the same name, calculate the distance between them, sort them from nearest to furthest, and then go down the list looking on Google Streetview until I found two houses with the same number. Easy.

So: I removed uniquely-named streets, and calculated the distance between any two streets with the same name. And I found that there were pairs of streets which were only a few metres apart. When I had a look on a map, I discovered that this appeared to be because if a street had a spur coming off it, that spur was sometimes listed as a separate street.

I also discovered that there were streets which had been fairly long in the past, but something had happened to split the street in two (like they built a park, or a school in the middle of it), and now the two ends of the street were listed as two different streets. Obviously, in neither of these cases would there be any house-numbers in common, so I needed to get rid of them.

I thought about just discarding any pairs of streets that were closer than a certain distance. Maybe 1km? Just to check, I had a look at a few random address-pairs around the 1 kilometre boundary, when I found this…

Image: Ordnance Survey/Google Street view/author provided.

Two streets with the same name, less than a kilometre apart, with at least one house-number in common! Blimey!

Um. So instead of discarding everything under a kilometre, I could actually discard everything over a kilometre. That meant a huge reduction in the amount of data.

This still left a lot of data though, and no way I could think of to procedurally remove the spur streets or split streets mentioned above, so I went through the data by hand.

It took ages.

This, by the way, showed up loads of errors in the Ordnance Survey data. The OS manage a vast amount of data and errors are bound to appear. If an error is of a type which is unlikely to be spotted during ‘normal’ usage of the data, it might never be found until some weirdo comes along, uses the data in an unorthodox way, and finds all the hidden flaws.

When I made my town & street name search apps, I’d already found dozens of errors in the OS data, but now I found loads more – especially duplicated streets. In one case, they had seven streets named Church Meadows within metres of each other. They turned out to all be the same street. Seven streets gives twenty-one pairs of streets – all erroneous. Aaaarrrgh!

The results

Yes! The bit you actually care about!

When I first wrote this blog post, I had here a top five of closest Address Twins, but for reasons I’ll get to later I’ve reduced it to just number one. The distance is front-door to front-door as measured with Google Earth’s ‘ruler’, rounded to the nearest five metres.

Image: Streetmap.co.uk.

2 George Street BB5 0HD, and 2 George Street BB5 0ET, both in Accrington, Lancashire, are only 235 metres apart.

Yes, that’s right. These twins are less than a sixth of a mile apart. When I set out to find the closest twins, I thought they might perhaps be around a mile and a half apart, which would itself be quite silly. But 235 metres is definitely in the “What were they thinking?” category.

Image: author provided.

It’s possible to stand at a point midway between the two houses where you can see both of them without moving. Sadly, you can’t quite see one house from the other.

But wait…

So, I was happy with that result. I wrote up this blog post, and I even visited the houses in Accrington and took some photos, while I was there for an exhibition.


But…

Before I posted this, I discovered that HM Land Registry make available data on their website about property sales in England and Wales. You can download all standard and additional price paid data transactions received at HM Land Registry from 1 January 1995 to the most current monthly data. So I got a copy to have a nosey.

I noticed that, amongst other things, the data contained the house number, street name, and postcode of almost every house that had been sold since 1995.

Now obviously this is not a complete list of every house in the UK, but because it was the nearest thing I’d found to a list of individual houses, I was intrigued as to whether it might show up any additional insights into the nearest address-twins.

Obviously there was a lot of data, and no geographic data which I could use directly to calculate distances between houses. So I generated a list of every pair of houses which had the same number and street name, and had different postcodes, but where the postcodes only differed by the last letter – because I guessed that very close address twins were likely to have very similar postcodes. The data was incomplete anyway, so I was just doing this out of curiosity.

And then, while checking out random house pairs, I found this…

Image: Streetmap.co.uk.

Um…

Oh.

I realised that one of my fundamental assumptions had been incorrect. I assumed that you couldn’t get two houses close together with the same number on the same street. Clearly you could. Those two houses are only about 130 metres apart.

Bugger.

So all the work I’d done up to this point was for nothing.

*quiet sobbing*

So now I had a choice. All the Ordnance Survey data was no longer useful. If I wanted to continue with the Land Registry data, then it would mean a lot more processing, because there was more data to start with, and none of it had any position data. Also, even if I found the closest address twins, I’d have no way of knowing if they were genuinely the closest in real life, because the data wasn’t a complete list of every house – as well as not including Scotland, it’s possible that there were entire streets where no house had been sold in the last 20 years. 

The Sunk Cost Fallacy

The Misconception: You make rational decisions based on the future value of objects, investments and experiences.

The Truth: Your decisions are tainted by the emotional investments you accumulate, and the more you invest in something the harder it becomes to abandon it. 

So… on we go then.

I removed duplicate addresses (if a house was sold twice, it appeared in the list twice, etc.). Then I removed flats (because they all share an address and just screw up the results). Then I found every pair of address twins. Then I grouped all the pairs by street. Then I geolocated the postcodes of every pair of streets. Then I calculated the distances and sorted the list by nearest to furthest.

This all took ages.

The resulting list still needed work because of a few factors. Firstly, the original file contained legacy data, so if a house changed postcode for some reason, it might be listed under both postcodes. Secondly, it appears that sometimes new-build houses are sold before they’ve had a postcode assigned, so the first time they’re sold they use a random nearby postcode – which means that, again, the same house might appear under two different postcodes.

And thirdly, some of the postcode geolocation data was wrong, because of course it was.

My enthusiasm was starting to flag by this point, so I thought that the first thing I’d do was to find the two nearest address twins which were in completely different postcode areas to each other. This involved processing much less data, and at least I’d have a result of some description – which would hopefully be an incentive to find the ultimate result.

The results (part 2)

What I didn’t expect was that almost immediately I’d find the ultimate result.

The holy grail.

Yes! The big one.

Image: author provided.

See these two houses? They’re both number 443 Manchester Road.

The one on the left is in Bolton; the one on the right is in Salford, both in Greater Manchester. They’re literally next door to each other.

Image: Streetmap.co.uk.

I cannot get over how ridiculous a situation this is. It’s possible to live in a house where your next-door neighbour on the same street has the same house number as you.

I spoke to a chap who lived on the other side of the road (who, quite rightly, wanted to know why I was taking photos of people’s houses). Apparently they’ve been trying for ages to get the two councils to put up signage to clarify the situation. You can see in the photo the sign on the Salford side which says “Manchester Road, Clifton”, but without a corresponding sign on the Bolton side, or maybe arrows to indicate that “Clifton” refers to the area to the right of the sign, it doesn’t really help.

So, the answer to the question “What’s the shortest distance between two houses with the same number and the same street-name?” is “no distance at all”.

PS…

You may have noticed that all the houses I’ve mentioned are in the North of England – particularly the North-West. This seems to be a trend. Of all the many twin-streets I looked at which I haven’t mentioned here, I’d say that at least 80 per cent of them were in the North, particularly the North-West. I’m not sure why this is.

My guess is that people used to have very localised lives, and it didn’t matter that lots of streets had the same names. The reason this situation remains in the North-West, I’m guessing, is that compared to other areas of the country, very little regeneration has taken place since then.

PPS…

It’s unlikely you’re thinking of actually visiting these houses, but I was nerdy enough to do so, so you might also, in which case remember that these are private houses. Don’t hassle the people who live there. Try not to look like you’re planning a robbery. You know, just be sensible about it.

This article originally appeared on Paul Plowman’s personal blog, and appears here with his permission.

Want more of this stuff? Follow CityMetric on Twitter or Facebook

 
 
 
 

Covid-19 is highlighting cities' unequal access to green space

In the UK, Londoners are most likely to rely on their local park for green space, and have the best access to parks. (Leon Neal/Getty Images)

As coronavirus lockdowns ease, people are flooding back to parks – but not everyone has easy access to green space in their city.

Statistics from Google show that park attendance in countries across the globe has shot up as people have been allowed to move around their cities again.

This is especially true in urban areas, where densely populated neighbourhoods limit the size of private green space – meaning residents have to go to the park to get in touch with nature. Readers from England can use our interactive tool below to find out how much green space people have access to in their area, and how it compares to the rest of the country.

 

Prime Minister Boris Johnson’s announcement Monday that people are allowed to mingle in parks and gardens with groups of up to six people was partially following what people were doing already.

Data from mobile phones show people have been returning to parks across the UK, and also across Europe, as weather improves and lockdown eases.

People have been returning to parks across the world

Stay-at-home requirements were eased in Italy on 4 May, which led to a flood of people returning to parks.

France eased restrictions on 1 May, and the UK eased up slightly on 13 May, allowing people to sit down in public places so long as they remain socially distanced.

Other countries have seen park attendance rise without major easing of lockdown – including Canada, Spain, and the US (although states there have individual rules and some have eased restrictions).

In some countries, people never really stopped going to parks.

Authorities in the Netherlands and Germany were not as strict as other countries about their citizens visiting local parks during lockdown, while Sweden has famously been avoiding placing many restrictions on people’s daily lives.


There is a growing body of evidence to suggest that access to green space has major benefits for public health.

A recent study by researchers at the University of Exeter found that spending time in the garden is linked to similar benefits for health and wellbeing as living in wealthy areas.

People with access to a private garden also had higher psychological wellbeing, and those with an outdoor space such as a yard were more likely to meet physical activity guidelines than those without access to outdoor space. 

Separate UK research has found that living with a regular view of a green space provides health benefits worth £300 per person per year.

Access is not shared equally, however, which has important implications for equality under lockdown, and the spread of disease.

Statistics from the UK show that one in eight households has no garden, making access to parks more important.

There is a geographic inequality here. Londoners, who have the least access to private gardens, are most likely to rely on their local park for green space, and have the best access to parks. 

However the high population in the capital means that on the whole, green space per person is lower – an issue for people living in densely populated cities everywhere.

There is also an occupational inequality.

Those on low pay – including in what are statistically classed as “semi-skilled” and “unskilled” manual occupations, casual workers and those who are unemployed – are almost three times as likely as those in managerial, administrative, professional occupations to be without a garden, meaning they rely more heavily on their local park.

Britain’s parks and fields are also at significant risk of development, according to new research by the Fields in Trust charity, which shows the number of people living further than a 10-minute walk from a public park rising by 5% over the next five years. That loss of green spaces is likely to impact disadvantaged communities the most, the researchers say.

This is borne out by looking at the parts of the country that have private gardens.

The least deprived areas have the largest gardens

Though the relationship is not crystal clear, it shows at the top end: Those living in the least deprived areas have the largest private green space.

Although the risk of catching coronavirus is lower outdoors, spending time in parks among other people is undoubtedly more risky when it comes to transmitting or catching the virus than spending time in your own outdoor space. 

Access to green space is therefore another example – along with the ability to work from home and death rates – of how the burden of the pandemic has not been equally shouldered by all.

Michael Goodier is a data reporter at New Statesman Media Group, and Josh Rayman is a graphics and data visualisation developer at New Statesman Media Group.