Dmitry Lebedev: “A few years ago, I clearly understood that a little bit more and OpenStreetMap would sink”

Dmitry Lebedev is a master of economics, programmer and urbanist who has been working with OpenStreetMap for over 10 years. He not only draws houses, but also does a lot of research based on his data. Which way OSM went, whether it has a future and why programmers need humanities - he told all this in an interview.

- How and when did you meet OpenStreetMap?

- I definitely can’t say it was more than 10 years ago. I was interested in the project, I heard about it that it was "like Wikipedia, but only about maps." However, I did not register immediately, I did not want to start another account on another service. But at some point, curiosity got the better and I started editing the map. First I drew the territory around my house, and then it dragged on so that for several years I was the first in the ranking of cartographers in the Novosibirsk region.

- What exactly interested you in the project?

- The fact that you almost immediately see the result of your work: you just drew, but it has already been displayed on the map. This is incredibly catchy. In addition, remember what Yandex and Google were 10 years ago? What can I say, even 2GIS at that time was not so detailed. And in OSM even then it was possible to draw everything you want: intra-quarter roads, trees, playgrounds and much more. Due to the fact that I cycled a lot, I quickly appreciated the OSM potential: I drew the roads I needed, poured into the navigator and went for a drive. Therefore, for a while I mapped a lot.

“Do you continue to map now?”

- No. Occasionally I will draw about a dozen houses and that’s all, for more there is already not enough patience. At some point, everything bothers, interests and priorities change, and personally it happened to me with OSM.

- What was the project then? How did you live and breathe? What has changed during this time?

- If in 2010 it was still possible to go to the Russian forum and get an adequate answer to your question, then in 2016 it is no longer there, as it gradually turned into a garbage can, where only aggressive trolls “sat”. I don’t know how things are now, but if a newcomer appeared, he was often “pecked” by more experienced participants, who for some reason thought that this person was asking stupid questions and consciously wanted to harm the project with his activity. Therefore, discussing anything in such a community has become meaningless.

Do you know what else has changed? Ten years ago, at OSM, it was possible to achieve something and promote some kind of initiative with enthusiasm. Now I don’t see this, and I think that this is completely unlikely. To change something in the project at the moment, you must be a full-time specialist in interacting with the OSM community in some IT company. Nothing wrong with that. This is a natural project development process. He grew up and major market players drew attention to him, with whom came not only money and politics, but, above all, the future.

Probably, it was not me who, several years ago, clearly understood that OSM would go down a bit. Then it seemed that if nothing had changed in the project, then his funeral was not far off. But business came into it and canceled them. So 10 years, and maybe even 20, the project will still live. It seems to me that he has now followed the Android OS development model, which has two segments in its development. The first is the Linux kernel, made by professional enthusiasts for themselves. The second is the commercial component aimed at the mass user.


Project " Tram systems of Russia "


Project " Dedicated lanes "

- As far as I know, you, as an analyst, often use OSM data in your research. Was it difficult for you to start working with them?

- No, since I was a participant in the project and drew a map myself. I saw it from the inside and understood what it consists of. Therefore, I had no difficulties. In general, OSM is simple enough. Another thing is that there is no single tagging scheme in it - who is what. But even you get used to it over time.

How do I usually do it? I begin my experiments with small district centers. Most often, there are no difficulties at this level. If the model works, then I'm trying to apply it to cities with a population of over one million. And now here sometimes there are errors in the data that prevent the algorithm from working. They have to be closed manually. For example, somehow I decided to calculate the population density by quarters based on OSM data, namely by the tag that indicates residential areas: landuse = residential. Everything was perfect, but in one place I had an incredible density, because someone dwelling an apartment building also surrounded the area with a tag with this tag.

About this nature, most often there are problems when working with OSM data. But, as it seems to me, I found some data types in OSM that you can safely take, for example, a road graph. Quite a lot of people follow its coherence and quality, and therefore there are rarely problems with it.


Dmitry Lebedev at the Party of the Revived Cards. Photo: Andrey Dremov. Presentation in PDF

- What tools do you use for working with geodata?

- I use GeoPandas, Osmium, GDAL, QGIS, for some time I worked with PostgreSQL, but after that I refused it. He is good in that you can quickly start working with him, but if you experiment a lot, he begins to interfere. There are still a number of my own scripts and developments, but so far I can’t share them.

- What would be advised to those who are just going to engage in data analysis?

- I think that in order to learn this, one must have before himself not a lecture, but a real task or a data set. And in order to know what to do with this, any overview course of statistics, econometrics (regression analysis), as well as some research presentations - to find out what people do with the data, will come in handy. Materials are full of such complete, and it is difficult to even remember something specific.

Several interesting authors talk about the principles of development in general: for example, Douglas Crockford, the author of the JSON format, gave 7 interesting lectures back in 2009, but those related to history ( 1 , 2 ) are interesting to listen to now. Jack Didrich, one of the developers of Python, gives a report “ Stop writing classes ” (I translated it), as well as a wonderful article by Matt Linderman on design speed.

Please pay attention to the last text. In short, it tells about a specific story about why it is worth doing many small iterations rather than spending months developing bulky models that eventually break in a minute. Achieve results through quick trial and error, not great failures. It is very important to remember when working as a programmer. Although, it seems to me, this principle can be useful in life as a whole.




Project Dmitry Lebedev " Beer against coffee "

- What do you like about OSM? I do not like? What would you do better?

- First of all, I would redo the data model. Which one should it be? I don’t know, but definitely not what she is now. For some reason, it seems to me that developers should look towards storing data in the form of versions. Let's just say a certain analogue of patches in git with the option of versioning and allocating clean branches. Is it possible to implement this with a geographic database, where an incredible amount of change? Question. But if IT giants came to OSM, then I would allocate a grant in their place for the development of a prototype of such a database, and it is important that it has topological structures, because, for working with data as in a database or Shape / GeoJSON / GPKG, no problem.

I would also pay attention to the existing “patch” - multipolygonswhich are now made as sets of lines. Because now you can open the multipolygon of some sea and fill the entire map with water. And try to find a place where it was opened ... On the STOSM I read ideas about this I don’t remember how many years ago.

I would also be in the place of the OSM Foundation , which has recently become preoccupied with diversity, considering diversity not only as creating the conditions for the project to have more women, people in Africa, South America or another continent on the planet. At the moment, any healthy person in the world has a hypothetical opportunity to become a member of OSM. But here who really is disconnected from the project is people with disabilities.

For example, we do not have a site version for visually impaired people - for them the map is poorly adapted. Or those with limited mobility - there is little data for them. It seems to me that OSM has the opportunity to open the door for such people tomorrow, as there is a whole technological basis for this.

You can start by making a separate card or card style for the visually impaired, then adapt the tools for them to create the map for them. And in itself this is a very important point - orientation in space of people who have vision problems. Can OSM help in this situation? Can we draw a card that takes into account their wishes and helps them every day? This is a real task, not a tribute to fashion.

Therefore, I believe that any conversation for inclusiveness should begin with the search for those whose help will give a tangible result. If diversity is considered only as a process - an endless struggle for it - then this struggle only discredits the one who leads it.



- What would you say to the person who is thinking: to use OSM data or not? To take part in the project or not?

- Definitely worth a try. Why not? This is the internet. If you look, study, draw a couple of houses, learn something new, you won’t lose anything. You should not be afraid of new knowledge and experience, as well as new data. I like the way they do in American startups: they boldly take and try to apply for their needs, spend a day or two on experiments and prototypes. If in the process they understand what does not work out, then they leave it.

Unfortunately, in Russian non-startups, especially in various near-state hard-to-pronounce institutions, and sometimes in outsourcing, they try to avoid any experiments. They are afraid of spoiling the working process and love obedience. If their position is close to you, then what can I say? Well, then do not try to work with OSM data, do not look in their direction, do not approach them.

People are often captives of what they grew up with. According to Crockford, for an IT innovation to come into use, it takes about twenty years for a generation of specialists who think in the old way to retire. For example, object-oriented programming was invented in the 60s of the last century, and it began to enter the circulation only in the 80s. And so with a lot. Crockford says that many in disputes over various innovations take the stand “try to convince me that your technology is needed, but I have not tried it and I do not want to learn new things”. If you do not reflect on your own views, then you can also get stuck in development.

OSM - 15 years. The project continues to evolve. And I think his time will come and he will be everywhere. Not so long ago, large players came to it. Even if they appreciated its potential, what other signal is needed?



- What have you learned thanks to OSM? What have you discovered for yourself?

- I learned quite well geography and place names, including my small homeland - Siberia. It turns out that we have a third of the Turkic names. OSM is also a good example of a non-profit IT project, on which you can analyze some aspects of game theory . It is one thing that I studied it at the university, and another thing - I saw it all with my own eyes live and from the inside.



- What do you say at the end of the conversation? Wish our readers?

- It so happened that most of them are programmers or people associated with IT, and much less often than humanities. This leaves its mark on how the project is developing, where it is moving and what is happening with it inside.

For example, IT people like to appeal to how everything is planned and how it should be in an ideal world. In urban planning and architecture, it has long become apparent that everything living does not develop as it was in the plans of the organizers: people all over the world do not walk on perpendicular sidewalks; for some reason, people don’t want to send their children to kindergarten or school inside their microdistrict; science cities, even during the Soviet years, did not produce massive scientific breakthroughs. The social and human sciences have long been studying why this happens. You don’t need to know about them at all in order to naively figure out how to organize, for example, a community, and wait for it to work this way.

There is an extreme example - researchers of terrorist networks, for example, Mark Sageman, or Diego Gambetta notethat they are made up not of dense fanatics, but well-educated idealists with a technical background and utopian dreams. Apparently, humanities, knowing history and previous similar teachings, are able to recognize dogmatism and manipulation.

Therefore, it is necessary to expand knowledge in the humanitarian sphere. For example, get acquainted with at least Popper’s theory of science . After all, all the programs and data that we are so fond of, for which we are ready to swear for days on the forums, are all an incredibly small part of real life.

PS Report of Dmitry Lebedev “Research with open data and GIS” ( video , presentation ) (“Open Data Day”, Moscow, 2018).

---

Communication of Russian OpenStreetMap participants is in the Telegram chat room and on the forum .
There are also groups on social networks VKontakte , Facebook , but they mainly publish news.

→ Join OSM!

---

Previous interviews:

, , , , , , , , , , wowik, SviMik, , , , , , aka BANO.notIT, , , , , .