Users' Diaries

Recent diary entries

Wikidata is a free knowledge base for linked open data designed to support Wikipedia and its sister projects, such as Wikivoyage. It contains over 97 million entries structured as a “Labeled Property Graph,” which is more powerful than RDF-based graphs. Like OpenStreetMap (OSM), Wikidata (WD) is an open crowdsourcing project with a large and active community.

Since 2014, OSM can be linked to WD through its tags. Currently, there are about 5.5 million such Wikidata tags with steadily growing popularity. These links can be used to create interesting products, for example a map with castles enriched with factual data from WD. However, the quality of these manually captured links in OSM is as yet unknown and untested. One must also note that the preferred way from WD to OSM - the other way around - is to use only coordinates (WD property P625) - i.e., no WD properties such as P402 are to be used because this covers only OSM relationships.

Now, two computer science students, Jari Elmer and Timon Erhart, from the Eastern Switzerland University of Applied Sciences (OST) have developed an application called “osm wikidata quality checker” with the aim to check the existing links from OSM to WD. The errors found - for example invalid WD entries in OSM - are also sent to osmose with a suggested correction. Osmose is a quality assurance tool for detecting problems in OSM data, which is also integrated into the iD editor, among other applications. The goal of the application was to become an integral part of OSM’s quality assurance ecosystem. It handles the large amounts of data in the two databases (about 1.5 TB each).

The successful result of the thesis is a data processing pipeline capable of finding diverse types of erroneous Wikidata links in OSM with a high accuracy of >95%. By using multiprocessing and the developed database model, where only the relevant data is extracted, the tool is able to handle the large amount of data and check the whole world on a weekly basis. The difficulties of dealing with crowdsourced data, where unforeseen data errors are to be expected, were also mastered, resulting in a robust software. Documentation and an easy-to-understand architecture allow the tool to be extended and additional checks to be implemented. The optional configuration provides the necessary flexibility in operation and helps with further development.

Currently, a total of over 30,000 errors are found in the following nine categories:

  1. Incorrect value for Wikidata-Tag
  2. Wikidata item does not exist
  3. Redirected value for Wikidata tag
  4. The distance between OSM object and linked Wikidata item is unusually large
  5. Characteristics of the OSM tags and linked Wikidata item do not match
  6. The secondary Wikidata tag and the linked Wikidata item do not match
  7. The OSM object is linked to an unpermitted Wikidata item
  8. Unpermitted link to an instace of living organism on Wikidata
  9. The OSM object does not match the Wikidata item

We are happy that these categories already have been incorporated into Osmose (see e.g. this Tweet). We are now searching for a permanent place to host this data processing pipeline.

Location: Rapperswil, Rapperswil-Jona, Wahlkreis See-Gaster, St. Gallen, 8640, Switzerland

Date - 07/06/22
Time Spent - 0.75 hrs Activity:
1. Trying to decide focus of next steps. Pedestrian footpaths or Lane mapping accuracy?
2. Tried to evaluate if small improvements to pedestrian footpaths can make it easier for kids/parents to walk to school?
Used to create a map_potentialwalkway.geojson trace to see where a footpath trace can be drawn. More to come about this exercise.

Location: Northwest Columbus, Columbus, Sharon, Franklin County, Ohio, 43235, United States
Posted by Harry Wood on 6 July 2022 in English (English). Last updated on 7 July 2022.

Next week we’re having an OSMLondon pub meet-up for the first time in a while. Or at least I am. People don’t seem to like setting themselves as “attending” on, but I don’t think I’ll be alone in the pub. Looking forward to it anyway!

Today I wanted to print out a map of “Parkland Walk”, a local nature trail (and former railway). This was a project for/with my 6 year old son, which I spent a bit longer on than I should have today. In his class they’re doing various activities related to Parkland Walk. I thought it would be fun to give him a big map in style which he could colour in.

parkland walk map printout

The nice thing about a style for colouring in, is that it can be no style at all! Just nice thin lines for every “way” from the OpenStreetMap data was all I needed. I knew that would give him lots of intricate little building shapes and an outline of the Parkland Walk area itself. I thought I’d try send vectors to print to get these lines looking nice and sharp, rather than my old crude technique of screenshotting JOSM wireframe view, and inverting the colours.

So I wanted a simple way to go from an OSM file to SVG. In the end I did it as follows:

  • Download the area in JOSM
  • Delete some trailing “tails” off the edge of the bbox.
  • Delete some unwanted way clutter: admin boundaries and underground features.
  • Save as… and set the file name extension to geojson.
  • Open in a browser
  • Drag the geojson file on there and “Import” with default settings.
  • Export and select SVG. default settings.
  • Open the resulting SVG in inkscape
  • Select the black background polygons
  • Go to the “Fill” styling options and set fill to none (“x”)
  • Go to the “Stroke paint” options and take it off none onto black colour.
  • Go to File->Document properties and set the page size to A4
  • Resize the whole image to be much bigger fitting to the A4 page as desired.
  • For my case I wanted three different pages to stick together, so I also sized and moved the image to give me the framing I wanted, then moved it again for each page one-at-a-time.
  • While resizing I also stretched the image vertically a bit to crudely compensate for losing projection information.
  • After making it all bigger, the line widths are relatively thinner which looks a lot better (unless you resize line widths too), but you may want to set the “stroke paint” line widths thinner still.
  • Add a little OpenStreetMap credit text in the corner.
  • Send to print! (Or save as PDF)

Here’s how it looked in inkscape at the end:

framing A4 print area in Inkscape

I tend to think in terms of SVG when it comes to vector printing, maybe because I enjoy being able to preview and post-edit things in inkscape. But perhaps I should learn to do these things in QGIS (and print direct from QGIS). Certainly seeing things like 30 day map challenge it became more clear to me that people do produce finalised polished map output directly in QGIS (or ArcGIS etc). And I see it has a clever print layout options. Admittedly I’ve only ever used QGIS for fuddling around and converting from shapefiles and maybe doing geo operations (calculating buffers etc).

I have used Maperative to create styled SVG in the past, and I would recommend it. Excellent SVG output with useful object groupings once you get it into inkscape. But is .NET stuff a bit messy to install on mac?

As mentioned, I’d decided I didn’t need any styling, and so installing either QGIS or Maperative seemed like overkill, but actually I did start looking at those basic lines for roads, and started thinking these could be more fun as areas to be coloured in (introduce a bit of styling).

Also the “simple conversion to SVG” wasn’t as simple as I’d hoped. Quite a few steps in my list above, buts that’s not counting all the experiments and dead ends I hit. I hadn’t used mapshaper before, but followed this answer to arrive at that. I did also follow the instructions to install and use osmtogeojson before realising that JOSM can simply save to geojson! So that’s something which turned out to be pleasingly easy. Could JOSM make that more obvious? A file type picker? A separate “Export as…” menu option? Also should JOSM be writing projection information into geojson? Or what option should I have used within mapshaper to specify projections? …Or could JOSM just save as SVG? Casual feature request :-)

There’s a quite a few different ways to create an SVG map from OpenStreetMap. I remember working on building that wiki list a long time ago, and I see it has been extended since. Please do edit if it’s missing any good options.

Anyway once the six year old got home, his first challenge was to stick the three sheets together, lining things up properly. After that I thought he’d launch into colouring in all the little buildings in a rainbow of different colours, but he’s been learning about maps and so he immediately said “We need a key” and went on to choose a few specific things he wanted to colour in (and add to the key). So it kept him entertained for half an hour. It kept me entertained quite a lot longer than that!

If you’d like to help keep me entertained… come to the pub with me! We’ll be in the Jack Horner on Tottenham Court Road next Tuesday evening. All the details.

Location: Crouch Hall Park, Crouch End, London Borough of Haringey, London, Greater London, England, N8 8HL, United Kingdom
Posted by b-unicycling on 6 July 2022 in English (English). Last updated on 7 July 2022.

I’m preparing a video tutorial on thatched buildings - it’s dead simple, but it’s an interesting topic, I thought. So I was doing a bit of research trying to learn something about local thatching traditions and came across a 1994 survey by - it turns out - a thatcher. I only came across one volume which is basically a photo album of 106 thatched buildings in Co. Kilkenny with handwritten captions given the location. Location being in most cases the townland. But because it is handwritten and because this is Ireland, I have not been able to identify all the townlands.

What I’m doing now is trying to find the townland and trying to identify the building. Most times, if there are other buildings in the picture, it is possible to spot them in the townland by comparing the arrangement of buildings. Luckily, most buildings in Co. Kilkenny are mapped thanks to our #osmIRL_buildings project. And luckily, those are mostly old buildings, so I don’t have to worry about them being built since 2019. The National Index of Architectural Heritage is somewhat helpful in that they have indexed some of those buildings, but not all. They have a map where you can find a blue dot for those marking the spot. However, they are not always correct. They also have pictures of those buildings (sometimes also the wrong ones) which I can compare to that survey/ photo album. Sometimes, rarely, because those are very rural areas, I have mapillary to work with.

If they are not too far away and accessible for a non-driver like me, I’ll go and check them out in situ, take mapillary and a photo or two for Wikimedia. #OpenData, baby!

You can kind of see on Esri World Imagery (Clarity) Beta, whether it is thatched or a slate roof, because the thatched roofs have smoother corners and the colour is a bit different.

When I think that I have positively identified one, I add roof:material and wherever I have that information, building:levels and roof:levels (only if I get a gable view or there are dormer windows) and ref:IE:niah and wikimedia and mapillary links. This will result in a much better machine readable database than NIAH, hopefully.

Future/ possible extra keys

I would like to also add thatch:material, but I can’t tell from the photos (yet), and it isn’t an established key (yet).

It would also be interesting for someone interested in vernacular architecture to record since when those buildings are thatched, because some are recently built or only recently re-thatched after having had a different roof material for decades or longer. They are not as much affected by the necessity to source the material locally as was the case a hundred or two hundred years ago.

Some are also no longer thatched; maybe I should use was:roof:material=thatch, even better with a life cycle tag.

What can be learned?

Even when trying to locate the townlands, I got the impression that the few remaining thatched buildings are concentrated along rivers. That makes sense, because that is where the material used to be grown. (Not any more thanks to climate change and pollution - it now has to be imported from Turkey.)

Side effects

  1. While trying to identify the building, I have come across unmapped buildings that were overlooked during the #osmIRL_building task for Kilkenny. Some are not mapped neatly, so I can go and fix that.

  2. While trying to locate the building on the NIAH map, I have found mistakes in their location where the alignment of buildings does not match their image and mine. I can report those to the NIAH, and hopefully, they will fix it. I think that they only survey every 10 years, so that is how long it can take them (if they discover that mistake) to fix it.

  3. I’m filling the newly created Wikicommons category Vernacular architecture in Ireland.


Here’s an overpass-turbo query of the buildings and the main rivers in Co. Kilkenny (I had to create a relation for one of them, too, and other rivers need some TLC).

I’m using the hashtag #thatch for those changesets.


Location: Knocktopher, The Municipal District of Callan — Thomastown, County Kilkenny, Leinster, Ireland

I have an endgame in mind, which is having a complete walkability study on a bunch of major cities in the world, which competes in quality with but it’s fully open data.

That’s too grand for me to accomplish in a year of less-than-part-time effort. So I’ve decided to scope it down to a single city I care about, which is Vancouver. How hard can it be to analyze walkability in a single city? Well, pretty damn hard actually.

The very first thing I’d like to consider analyzing walkability is proximity to amenities like cafés, markets and pharmacies. Turns out OSM seems to have pretty good coverage on shops in Vancouver already, but for me to be very confident on my analysis the study stars with an evaluation of data coverage.

If I’m going to compare OSM with an official source, say Vancouver’s Storefront Inventory, whatever the coverage might be… I might as well import what’s missing? I think I owe OSM this much, and it will be nice to say that the whole data used in the walkability study is from OSM instead of from multiple sources.

The thing is, my past experience with data imports is limited to a single one I made for Wikidata of Higher Education Institutions in Brazil, and it took a whole month to finish it. I had more tooling available (OpenRefine is very integrated with Wikidata), more time available (5 years ago I had more energy) and the data was much more straightforward (no mapping involved, just categorization)…

Considering all that, I’d estimate it would take me about 3 months to import the whole thing following proper procedures. So I’ve decided to scope it down once again, to a subset of the data that I can reasonably scan through manually. I’ll start with coffee shops / cafés only. That I think should bring the estimate back down to a single month or so. Hopefully.

In my opinion, one should add to Openstreetmap functions to display the visibilty of astronomical events, like Solar eclipses. Corresponding ideas should be made at

Date - 07/05/2022
Time Spent - 1.5hrs
Changeset - #123254900
1. Focused on improving junction and road definition at Bethel Road & Pickforde Dr.
2. Split Bethel Road and Pickforde Dr. by adding nodes.
3. Added lane information for Bethel Road and Pickforde Dr. going into and out of intersection.
4. Reviewed turn assignments at junction. They look okay.
It was difficult to understand the correct approach to mapping a junction. The OSM Wiki documentation can benefit from additional examples and best practices.

Location: Northwest Columbus, Columbus, Sharon, Franklin County, Ohio, 43235, United States
Posted by Andromeda 333 on 5 July 2022 in Russian (Русский).

Я впервые здесь. И не совсем понимаю, что нужен делать. Точнее, совсем не понимаю. Объясните пожалуйста кто-нибудь)

Posted by mcvillain on 5 July 2022 in English (English).

Dear Diary.

I made my first edits today.

  • mcvillain

In my last diary entry I wrote about putting the requested note ids into url. I said that it’s likely not going to be a problem because ids are stored in the url part that’s not sent over HTTP therefore it doesn’t matter how long is this part, how many ids it contains. That was said about a typical output of note feeds for countries that contain updates dating back up to a week. For a busy country that’s usually a few hundred notes. Concatenated with single-character separators they result in a few thousand character string appended to a url.

Feeds are not the only way to get note ids from In fact you may no even want to get notes from the feeds because of their one week age limit. There are regular html pages for each country too, one for all notes, another for open notes. They likely contain more notes than the feeds do. The open notes pages, which are of interest to someone who wants to resolve notes in a given country, can have up to ten thousand notes. I added options to open those with note-viewer, along with arbitrary html files. The last two options in the Get … notes for this country dropdown of the XML tab set up the selector (cells in second columns of tables from webpages) and open the webpage that you have to save to a file and then open that file with note-viewer. Unlike the xml feeds, even Firefox won’t start the download automatically.

Is note-viewer able to open ten thousand notes? Kind of. It won’t have to load all of them at once. There’s not even a way to load them at once. Like was said in my last diary entry, each note has to be downloaded with a separate request to the OSM API. That’s why the available batch sizes in note-viewer’s XML mode are so small. Even if there was an API call to fetch several notes at once, like the one for elements, it wouldn’t work for thousands of elements. In case of such hypothetical API call, the ids would have to be sent over HTTP, this is going to limit the request length. But let’s go back to urls generated by note-viewer that can supposedly be of any length. Will note-viewer generate a url with 10k note ids? Yes. We’re at seven-digit note ids now, plus one separator character between ids - that’ 80k characters! 80k characters in your browser’s location bar, will they fit there? Let’s find out:

Link example

“Latest changes” ( got a new update:

A tag comparison table opens when clicking on any element (similar to achavi).

In a previous blog post I explained what this tool is about and how to use it.

Multi devices

Yesterday, 4th It was so bad that dust easily settled on a lot of things, when I held out my phone for a minute or two, screen was covered in a thin layer of dust

After telling that no mapping for me today, by nightfall, it piped down and I could resume, now that I think about it, I don’t have anytime to be resting, gotta do the cycle everyday, there is way too much to do..

Funny enough (joke intended), I’m only doing single node POIs(shops, stores etc), not counting other stuff that is still missing, I kinda can’t wrap my head around the number of missing things from map.. Nameless alleys, streets and also those which are not drawn (there are some that don’t even have a sign and you have to ask locals to how to address it or if they know the name , since municipality website itself is incomplete) , missing buildings (apartments for example).. The list could go on and on

But and but, one thing at a time.. I plan to do a great deal of work till the end of summer , as much as I can and situation allows me..

Time will tell(probably)

Tout d’abord je tiens à informer qu’il faut vérifier sur le terrain souvent car c’est en travaux continuellement - Ajouter les éléments temporaires (parkings, broussailleset routes) - Ajouter les bâtiments en construction (immeubles, écoles, commeces et services publics) - Mettre à jour les transports (Star et Breizh go) - Guetter les potentiels travaux sur Enfin, Merci beaucoup aux contributeurs de la plaine de Baud.

Создал трек межну смотровыми площадками 57.93657, 55.88472 Смотровая площадка 1 Смотровая площадка #4730463780 57.93652, 55.88628 Смотровая площадка 2 Смотровая площадка #4730463783

Location: Нижние Муллы, Култаевское сельское поселение, Пермский район, Пермский край, Приволжский федеральный округ, 614521, Россия

Have you ever been given weird looks by people, or imagining people are glancing over you when mapping a poi, but in reality there is less likely for someone to watch you? I know, it was a long, boring and hard question to follow.. For mapping non existent poi(s), like, let’s say shops, mostly you have to personally go to that place. Seeing people frantically passing, each living their lives, whatever that may be. Standing there, sometimes get that feeling of being watched over, maybe you’re adding phone info off the sign of shop and owner is standing outside, looking at you like ““wonder what’s that guy doing over there?”

For me, a lot of times, I can’t just simply walk to the owner and be like “what’s your shop’s phone number?” or “what are your opening hour?”, specially if it’s not written on the sign, owner thinks like “is he/she a marketer or something?” or maybe he doesn’t.. Maybe if I were to ask such questions, now I have to explain them what is osm, which is nice and I would love to do so(assuming they wouldn’t know), maybe he be like “is his stupid, why waste time like this?” And I would selfishly blurt out “I’m doing this for people like you”

Am I enjoying this? Or maybe I’m deceiving myself

Maybe I’m being too sensitive and overreacting..

The sense of conflict in my own life has tangled up my beliefs about what am i doing as a contributor to OSM

but I sigh in disbelief, knowing such scenarios are all in my head!

Seit 2021 bin ich in der Region unterwegs und trage Daten zu Bildstöcken, Wegkreuzen, Marterl zusammen. Und die werd ich versuchen hier festzuhalten. Da openstreetmap vom Funktionsumfang weniger hergibt wie die Fülle an Info zu den christlichen Flurdenkmälern, beschränke ich mich auf die Positionsdaten und die Unterscheidung Bildstock - Wegkreuz. Marterl am Straßenrand, welche auf Unglücke hinweisen, werde ich im ersten Schritt nicht anführen. Ebenso muss ich erst herausfinden, inwieweit ich Bildbäume in der Karte festhalten kann.

On peut classer les chemins de randonnée par niveau de difficulté (et d’équipement requis) en utilisant l’attribut “sac_scale” :

  • sac_scale=hiking (randonnée)
  • sac_scale=mountain_hiking (randonnée en montagne)
  • sac_scale=demanding_mountain_hiking (randonnée en montagne exigeante)
  • sac_scale=alpine_hiking (randonnée alpine)
  • sac_scale=demanding_alpine_hiking (randonnée alpine exigeante)
  • sac_scale=difficult_alpine_hiking (randonnée alpine difficile)

Voir le détail sur : Key:sac_scale

On peut retrouver tous les attributs relatifs aux chemins de rando sur le wiki Randonnée pédestre

Con il 1 luglio 2022 la Trieste Trasporti (con supervisione di TPL FVG) ha deciso di rinumerare le corse della linea 6 tra Piazza Guglielmo Oberdan ed il Bivio di Miramare co il numero barrato per non creare confusione. Ho proceduto alla rinumerazione aggiungendo la /, è creato anche il nuovo route_master.

Ricordo che le corse deviate per Porto Vecchio, a sostituzione delle linee 80 (soppressa) ed 81 (sospesa nel servizio estivo) non hanno diversa numerazione, tantomeno capolinea rispetto alla 6 “normale”; allego foto scattata da me il 01/07/2022.

6/ in transito presso la pineta di BarcolaCon il primo luglio entra anche nuovamente in servizio la linea 90 tra Alto Adriatico e Lazzaretto, a rinforzo della linea 7.

Location: Roiano-Gretta-Barcola-Cologna-Scorcola, Trieste, Friuli Venezia Giulia, 34121-34151, Italia
Posted by ImmaBeMe on 2 July 2022 in English (English).

Summer came and with it, my work as an osm contributor became a lot harder, now I have to wait for sundown so weather is a bit more tolerable, I could use some help, but oh well, I guess it’s a lot to ask.. It seems not enough people contribute to map, at least where I live, even though myself isn’t doing that much, a lot others are doing almost nothing. Sometimes I see some activity here or there, they are being contributed via local map applications like Neshan or Balad(which is nice, but it seems as if it’s not actually osm) It’s noteworthy that there are some features you could only find in such apps, since they are designed to fit needs of our country as in map application aspect. Like likelihood of police presence, estimate of traffic, covid 19 map etc.

Sometimes I feel like I’m carrying a big burden on my shoulders, even though mapping supposed to be fun… Hmmm It’s not like anyone is going to read this anyway, so it doesn’t matter?

Notes of countries

One of possible uses of note-viewer is browsing notes in some area. You can do this using the BBox dialog that lets you get notes in a rectangular area. This is similar to how the note layer on the OSM website work and how the editors like JOSM load notes. And it similarly won’t work if the area is too large or not rectangular. So if you want to get notes in a country, you won’t be able to get them easily. But you may know that there are services that show notes for a given country, like the one by Pascal Neis. There you can select a country and get a list of notes filtered by their status.

That’s how you can get a list of notes for a country. But what if you want to see notes on a map? The list doesn’t contain note coordinates so it can’t be done directly. The situation is similar to looking at a list of user’s notes on the OSM website, something that note-viewer was originally written to take care of. Can it handle country lists now? Actually it can handle any list with note ids. In addition to showing webpages with note lists, Pascal Neis’ site also has Atom feeds. It’s going to be easier for us to get note ids from these feeds. And no, unfortunately, they still don’t contain coordinates.

Any list in our case is an XML file, and note-viewer handling it means that it’s possible to come up with selectors for XML elements that contain note ids. If ids are inside some attributes, the names of the attributes also need to be specified. Click on the XML tab in note-viewer and look below the instructions for to see the inputs for selectors and attributes. You won’t have to use them if you only want to get note lists from, just follow the instructions above.

Getting the list

The instructions seem a bit long. What you’d expect to do is to pick a country, pick a note status such as open or closed, press a button and see the notes in note-viewer. In fact, this is what you do except note-viewer can’t read the feed directly so you won’t see any notes yet. What you’ll see is either the feed file downloading or a new browser tab showing you the feed contents. In the latter case you need to save it yourself. All of this is because of same-origin policy that stops scripts from accessing data across domains unless the server permits it. server doesn’t.

Hopefully it’s not a big problem, just an extra file save and load, like was discussed in my diary entry about getting iD to open notes exported by note-viewer. And the solution is similar: drag and drop, except note-viewer now is on the receiving side. Find the file in the download manager (toolbar button in Firefox, bar under the webpage in Chrome) and drag it to the Read XML file area below the form in note-viewer. This will get the notes into note-viewer, although they’ll load slower than usual because each one of them has to be requested from the OSM API separately.


Normally, after you press the Load button, note-viewer stores request parameters in the webpage url. This way the request can be repeated if outdated or if opened on another computer. But in case of country feed requests its impossible to do because they can’t be automated. If you want a note-viewer url corresponding to open notes in some country, you are not going to get it. What note-viewer does instead is it saves the exact set of note ids that were present in the feed file when you opened it. If you reload the page you’ll get the same set even if the feed has changed. If you want new notes you’ll have to reload the feed.

Saving every note id in the url makes it rather long. There are ways to reduce its size, something that I’ll maybe do later. First, doing it will make the url unreadable. Also, its length may not be a problem. The length spent on encoding note ids is not subject to HTTP protocol limitations because ids are contained inside the fragment part after #.