Imagico.de

blog

October 28, 2019
by chris
0 comments

Satellite image news

The ESA recently made an interesting announcement about extending Sentinel-2B image recordings to lower sun positions in the northern hemisphere winter.

A bit of background on that: For the past years Sentinel-2 image recordings have traditionally been made only with relatively high sun positions compared to Landsat 8. The difference in the routine latitude based recording limit was not huge but it was noticable. This has led in particular to a significant under-use of the Sentinel-2 recording capacity in the northern hemisphere winter, especially since at the same time the Antarctic is only recorded at a lower frequency.

previous Sentinel-2 coverage in late December

Landsat 8 coverage in late December

The change in recording plans announced now is somewhat half-hearted – it only applies to Sentinel-2B and it so far seems to be limited to Europe and Greenland. Here is a low sun position image recorded with this new pattern.

Low sun position Sentinel-2 recording in eastern Greenland

More generally speaking the recording patterns of Sentinel-2 are fairly static based on a fixed and politically decided recording plan while Landsat recordings are based on a priority list and typically they try to record as many images as possible given the operational capacity available. Or in other words: Sentinel-2 could probably be recoding significantly more imagery, in particular on the southern hemisphere and of low latitude islands during the northern hemisphere winter, if the option to record these when capacity is available was worked into the acquisition plan. But that would of course require the political decision to do so.

The real question is of course why the sun position based recording limit for Sentinel-3 OLCI data is even tighter than for Sentinel-2 – here the northern recording limit on the same date as the above Sentinel-2 sample – which is several hundred kilometers to the south.

Sentinel-3 OLCI recording limit in eastern Greenland on October 25

Although Sentinel-3 records an earlier time slot than Sentinel-2 (10:00 vs. 10:30) and due to the wider field of view includes lower sun positions on the western side the sensible recording strategy would be to record everything that meets whatever sun position requirement is considered reasonable even if that practically means also recording positions with a lower sun position while in reality it seems no OLCI data is recorded where at the western end of the recording strip the sun position would be less than five degree above the horizon and this way you loose quite a lot of potentially useful data further to the east. In this case at the eastern end the recording starts at a sun position of about 11 degrees above the horizon.

Autum colors near Irkutsk, Russia in September 2019 recorded by Sentinel-2

GCOM-C SGLI images

Not exactly news but i wanted to mention that images from the Japanese GCOM-C satellite’s SGLI sensor are available now (and have been for quite some time apparently) on the JAXA data portal. GCOM-C is a satellite system somewhat similar to Sentinel-3 in scope but offers some interesting additional features like polarized light sensors and an ultraviolet spectral band. The data is available under a liberal open data policy.

Here is a sample image of southeastern Europe:

GCOM-C SGLI recording example

October 25, 2019
by chris
0 comments

State of the Map 2019 – visitor statistics

When writing my report on the State of the Map conference in Heidelberg this year i complained not having any data on the visitors yet. We now have some numbers and i would like to add the missing illustration and commentary on those.

SotM 2019 visitors – where they came from

Might be a bit difficult to read – i grouped the numbers by continent – Europe in yellow/orange, America in blue, Asia/Pacific in green and Africa/Middle East in red. I was a bit surpried by the large number of visitors from Germany – but if you keep in mind that this includes all the local helpers it might not be so surprising after all.

I prepared a second illustration with estimates for the carbon dioxide emissions generated from travel to and from the conference alone. This is a very conservative rough estimate. So don’t give too much on the exact numbers, neither individually nor in sum. I am pretty sure the actual emissions are not lower but they could definitely be much higher.

SotM 2019 estimated CO2 emissions due to travel (in metric tons)

Why am i showing this? First of all to show that a conference like SotM is a resource intensive endeavor. Second: To show that to reduce the environmental footprint of an event like this it is very beneficial to hold it close to where the majority of event visitors come from. Europeans this year have accounted for more than 2/3 of the conference visitors – yet they have probably contributed less than ten percent of the total CO2 emissions due to travel of the whole conference. On a per person level this conference – with an estimated < 600kg CO2 per visitor for travel - probably fares pretty well. And third: To illustrate that for the OSM community to be sustainable in the long term we will have to put a lot more effort into improving our ability to communicate and cooperate globally without the need to necessarily meet in person. We need to put a lot more energy into ensuring that we come to a point where physical presence at an OSM conference becomes truly optional and where you can productively contribute to such events - both actively and passively - at the distance.

October 9, 2019
by chris
8 Comments

State of the Map 2019 Heidelberg impressions and thoughts

It has been more than two weeks since SotM in Heidelberg and some are probably already wondering about my commentary. Part of the delay is due to me being busy with other things, part is due to some information i would like to have had not being available so far (i will get to that later).

The venue

Overall it was a pleasant experience for me. We had mostly luck with the weather so the choice of time of year was not too bad. I mentioned already last year that Heidelberg is conveniently close for me with only a two hours train ride to get there. Heidelberg – of all possible locations you could have chosen in Germany – is more near the upper end in terms of accomodation costs. But since a public transport ticket was included in the conference ticket for the duration of the conference for the whole regional transport area you were not limited to stay in Heidelberg actually so there was a pretty broad range of cheap options to stay – although probably not that easy to find for visitors from abroad. The information on budget accomodations on the SotM website could have been better – but that is a problem we essentially already had in previous years. Side note: I find it kind of annoying that the website has removed much of the pre-conference information after the conference – this is IMO not good style – maintaining previous information for future reference is important. Update: This has meanwhile been fixed – see comment below.

The venue itself was pretty well suited for the conference i think. All the main rooms for the talks were very close together – the rooms for the BoF sessions were a bit further away but quickly reachable as well. The most serious issue was IMO the acoustics in the large lecure hall (Großer Hörsaal) where – as you can see in the videos – the speakers frequently had trouble understanding questions from the audience. This is obviously hard to get right in such a large room and lecture halls like this are of course not really designed for dialogue between the speaker and the audience.

Conference size

Though we have no numbers and statistics so far for the conference visitors it was quite clearly the largest SotM conference so far. When the original planned number of tickets was sold the conference organizers increased the capacity beyond the original planning to allow more people interested to visit. And all of this together showed at various places during the event. Most obviously at the social event on Saturday where the catering turned out to be under-dimensioned – both in terms of amount of food and distribution capacity. Thanks to the good weather and the possibility to go outside the venue for the social event itself was fine – though a bit short in sitting opportunities outside the main room intended for eating, which turned out to be a bit stuffy and very loud – the whole venue was a former industrial building which in terms of acoustics was obviously not designed for a large number of people.

The poster session on Sunday evening was also a bit sub-optimal because the poster display and catering were separate on different levels so there was not really a natural looking at and talking about the posters while eating and drinking. This was also owed to the limited space mostly – even as it was the area where to drink and eat was crowded and there would have been no way to also appropriately display the posters directly there.

At the conference itself i did not see any problems with overcrowding or overfull lecture halls – at lunchtime at least during the first day the lines were long but there was plenty of space so it was not an issue.

In total my conclusion is that this kind of conference size exceeds the limits of what can be reliably managed the way SotM conferences have traditionally been operated. Specifically

  • organization at this scale IMO either requires a well-established team with multiple years of experience with a presence on-site during most of the planning phase or help from a professional event organizer with experience with this kind of event.
  • any location for a conference of this size with sufficient capacity – no matter if the main conference venue or for a social event – is going to be hard to find and organize so for a conference of this size the confirmed availability of all needed locations would usually need to be a pre-requisite for selecting a location for the conference – something that for SotM has so far not usually been a criterion (applications usually mentioned possibilities but rarely came with a full set of places with confirmed availability).

All of this is not meant to say the organizers did not do a good job – the opposite is the case IMO: Given the number of visitors it really went quite well.

The talks

I have not yet watched all the videos of the talks i did not see at the conference so here only a few select comments on the talks i recommended before the conference and others i went to.

  • Introduction to OSM: How it’s made and how it’s used (video) – although i have not seen this live at the conference based on the video this really worked out fine. This is definitely a format that could be built on in the future, potentially also for more specialized topics.
  • Communication and Knowledge Transfer in OSM (video) – as i hoped for this gave you a fairly broad look at the different communication methods used by the OSM community and their advantages and disadvantages. Really recommended for anyone who wants to take a look over the limits of their horizon of what channels and platform they are famiiliar with.
  • Mapathon, mapathon, mapathon! (video) – This offered an important critical view on common practices in remote humanitarian mapping efforts and their historic development. In my pre-conference post i likewise mentioned the followup talk (video) which focused more on organizational aspects of local communities in countries outside Europe and North America – which i can also recommend.
  • I specifically also want to mentioned a later event organized also by Nicolas and Severin – the Bilingual Breakout Session – Community building and empowerment in South: French-speaking countries in Africa+Haiti (video). I think this could nicely serve as a blueprint for future cross language communication formats at OSM related conferences. Language barriers are one of the main limitations for cross cultural communication within the OSM community and this format shows how this can be overcome with limited effort and lead to much improved communication between people speaking different languages.
  • Is the OSM data model creaking? (video) – this was a pretty solid analysis of the various practical problems when working with the current data in OSM for roads and paths for navigation purposes. Unfortunately in the end the considerations on how to address these problems were primarily data user centered and not mapper centered – in other words they wondered what might be the most convenient way to represent things in data form for the data user rather than for the mapper.
  • New processes to agree on tagging suggestions and their interaction with the editing software available on openstreetmap.org – considering the controvertial nature of the subject the discussion was actually pretty civil and meaningful. Roland already posted a summary of the results of the discussions and i really hope people will follow up on the ideas that have been discussed there.
  • OSM Vector Tiles in custom coordinate systems (video) – this was a very disappointing talk. The only thing they actually showed was maps in equirectangular projection – which, coming from Mercator, is kind of inverse evolution. No substantial discussion of any of the actual problems and challenges when creating digital maps in projections other than Mercator.
  • Board + Working Groups meeting (video) – this was rather interesting regarding the dynamics within the OSMF though ultimately not really that productive, which is how it was probably kind of expected by everyone. The format worked out quite well, there was a lot of commentary and discussion happening. There was quite a bit of what i would call essentially the OSMF circling around itself without much connection with the OSM community outside the OSMF but there were also plenty of interesting and valuable comments by various people i would encourage everyone interested in the OSMF to look up in the video (and if you were there – maybe re-contemplate them again).

Audience

I would have really liked to look at the number of visitors at the conference from different parts of the world here as i did last year but so far unfortunately no such information has been made publicly available.

My impression was that the audience composition was similar to last year in Milano with two main differences:

  1. there was obviously a significantly larger fraction of visitors from Germany.
  2. it seemed also that due to the HOT summit having taken place directly before SotM in Heidelberg there was a larger fraction of visitors with a HOT background.

One thing i noticed though is that although Germany has a very large local hobby mapper community – possibly the largest one world wide – there were relatively few pure hobby mappers at the conference. There is a huge overlap between the German visitors of this year’s SotM and the regular visitors of the FOSSGIS conference with an OSM background. Keep in mind though that FOSSGIS is not purely an OSM conference and FOSSGIS visitors with an OSM background are only a small subset of the German OSM community – despite FOSSGIS having free admission for active community members. There was some IMO quite understandable critique from the German mapper community that even the early bird community ticket price is kind of steep for a hobbyist. Given the conference was sold out quite early this understandably felt a bit like this financial barrier served at least partly to give professional visitors priority over local hobby mappers.

And i think everyone should be able to relate to a hobby mapper who has invested possibly thousands of hours into mapping their local area over the past decade and who therefore does not feel right about the need to pay EUR 75 for being able to visit a conference where others who largely have invested much less get their travel expenses paid in full.

Scholarships

Which brings me to the SotM scholarship program. I wrote about this before the conference but back then had only very limited data on just 10 scholars. Now we have a bit more information – both on the scholars and the selection process.

Some might wonder why i make so much fuzz about the scholarship program. The reason is that this is about quite a lot of money – in 2018 this was more than 20k GBP and given the larger number of scholars this is probably even more this year. While this is perfectly affordable for the OSMF that does not mean it is all right to spend all this money without proper accountability and consideration – money that could obviously also be spend for other things where it might do more good for the OpenStreetMap project.

Where OSMF SotM scholars 2017-2019 came from

Here is the updated map of where the scholars come from – this year as well as in the previous years. My analysis of this has not changed much – there are additional scholars from North America, Europe and (partly French speaking) West Africa. But the complete gap in Northern Africa, the Middle East and Central Asia is still there.

What we now have is – for the first time – numbers on the regional distribution of applications. These i tried to illustrate in the following maps. The first is for the total number of applications and the second is after filtering out formally incomplete applications.

Total number of applications per country

After filtering based on formal criteria of completeness of applications

I get two main observations from these:

  • the bias towards English speaking countries and former British and US colonies is even larger in the applications. More generally speaking much of the bias observed in where scholars came from in the last three years seem to be already pre-defined in the applications. This however does not mean the scholarship program can’t do anything about it. If the call for applications selectively speaks to people from some countries but not from others there is a reason for that.
  • there seem to have been essentially mass applications from a number of countries (Nigeria, Uganda, Ghana and Kenya with each more the 35 applications) and from none of these countries there was ultimately a scholar selected.

What we also have is more detailed documentation of the selection process – most of which however seemed to have not worked out as planned. A number of observations on the selection:

  • We have a list of people involved with scoring – however the documentation also said the plans for scoring “did not work out as initially planned”. We have no information on the scoring itself (like for example the anonymized scoring data in comparison to the final selection) See addition below.
  • There is a list of suggested criteria for scoring – however these are fairly vague and in parts also somewhat questionable. In particular the “unique story or experience to share” is highly prone to depend on cultural commonalities between the applicant and the reviewer.
  • The list of people involved in scoring lists 13 people – 8 of which have been scholars in 2018/2017. While i mentioned in last year’s comments that involving previous scholars in application evaluation could be helpful i also mentioned that disqualifying them permanently for applying for scholarships in the future would be absolutely essential for this to work. Otherwise you’d have a high probability of a ‘revolving door’ system evolving with people switching roles from scholar to reviewer and back every year. When you look at where the former scholars who scored applications came from (Kenya, Lesotho, Philippines, Germany, Russia, Uganda, Niger, Nepal) there is no clear pattern – while scholars were accepted from the Lesotho, Philippines and Nepal there were none accepted from Kenya and Uganda despite a large number of applications. So i don’t see indications for any actual impropriety in the process here but none the less in this form the system runs a very high risk of favouritism.
  • There seems to be no formal conflict of interest management of any kind. Given that quite a few of both the people involved in scoring as well as the accepted scholars have a job with some OSM connection or a formal position in an organization with OSM connections (like being HOT voting members) this is a reason for concern.
  • As i analyzed before the ultimate selection of scholars seems to be based on ensemble optimization rather than independent rating of the individual application and selection of “the top 20”. And as i read the documentation this final selection was done without any oversight by a single person who essentially decided where more than 20k GBP will go. If i was an OSMF financial auditor (and i am really glad i am not) this would be something i could not accept.

Note the obvious derivation between the social structure of the OSM community and the selection of scholars has a high likeliness of becoming a self reproducing system – even without former scholars being involved in the selection. Given the de facto preference for people with a job somehow connected to OSM or a formal position in some organization local hobby mappers from all over the world without such connections will realize that their chances for actually getting a scholarship are very small and will depend on them presenting themselves as being alike and compatible to the professional OSM environment. Introverts or people from cultures with predominantly different communication styles will have almost no chances because they do not match the established ideal for a SotM-Scholar.

Overall i think independent of the future of SotM the OSMF board needs to pull the plug on the scholarship program in its current form. Even if for the moment i give everyone involved the benefit of the doubt and assume that they all tried their best to accomplish a just and unbiased selection this is just a disaster waiting to happen – either through deliberate favoritism and corruption or just through plain incompetence. For neither scenario there seem to be mechanisms in place that would prevent this.

Any kind of scholarship or financial support program (and i specifically also have in mind ideas like microgrants here – which unfortunately might be destined to be managed in a similar fashion) would in my opinion need to be managed by and with broad support from the hobby mapper community. The fact that hardly anyone from the community seems to be currently interested in helping with that should tell the OSMF board and the SotM-WG that there is something seriously wrong with how it is run at the moment. Even if (or more precisely especially if) you feel scholarships are important to have you should work towards giving it a fresh start with a proper mandate and a solid ethical and procedural framework. This would give potential volunteers the confidence that they can be comfortable in contributing to something beneficial for the project. The more detailed documentation we have this year about how the selection process actually took place is much appreciated but to me this mostly better illustrates the lack of and the serious need for a proper framework of binding rules and control mechanisms.

Addition: There was more data and information added after i started writing this post. Everyone is encouraged to have a look at this.

The future of SotM

Coming to the last part and a renewed critical view of the idea of the SotM conference in total. I have expressed my concerns about this before, in particular about the illusion of SotM being a conference for the whole OSM community. Next year’s conference is now planned to take place in Capetown, South Africa. This makes it relatively easy for me since there is very little chance i would want to go there – at least not on my own expenses. Roughly estimated for the cost of visiting SotM in Capetown i could probably visit most of the local OSM conferences in Europe during that year which would allow me to meet more people and it seems overall more attractive than a single visit in South Africa.

When i originally suggested to stop having a dedicated international SotM conference and instead having the OSMF every year give special support to a regional OSM conference most reactions i got were negative. This year talking to people at SotM about the future of the conference i heard a lot more people essentially agreeing with that idea. Overall i would say there are two potential futures for SotM:

  • giving up the idea of an international SotM and instead giving rotating support to local/regional conferences fully managed and organized by the local mapper communities. The goal could for example be for the OSMF to provide financial support (through either OSMF funds or by organizing sponsorships) that allows the conference to offer free entry to local community members and thereby ensuring broad accessibility for local mappers. In addition the OSMF could organize video recording and live transmission of the events at the conference and this way facilitate broader reach and participation without the need for expensive and resource intensive travel.
  • giving up the pretense of SotM actually being a community conference and concentrating it on what it mainly is right now: A meeting of professionals with OSM connection and the international OSM jet set in addition inviting some locals of the OSM community from the place they meet at.

In the discussing for next year’s place for SotM (the decision was made last minute during the conference this year) the selection was between two applications – Rapperswil and Capetown – which even without the travel costs of getting there are both on the expensive side. The documented criteria for selection essentially already make it abundantly clear that this is not a community conference but primarily targeted at the interests of business visitors and wealthy cosmopolitan hobbyists. Affordability of a visit is nowhere to be found on the list. So in a way we are already pretty far into the second option. But broader realization within the community that this is the case – and next year’s decision for Capetown further underlines that – could also create more support for the first variant.

And just in case anyone wonders – this comment probably would have been more or less the same if the decision had been made for Rapperswil. My personal travel costs for a visit to Rapperswil would obviously be less than for Capetown but that is only because of the costs of the flight. The local costs would probably be even higher in Rapperswil. For most people from outside Europe the difference would probably be fairly low.

October 9, 2019
by chris
0 comments

How to not do Geo-Visualization

There has been a fairly impressive fail in geo-visualization in discussion on digital channels in Germany during the last days i wanted to comment on here.

It is the new banner image of the Twitter account of the conservative party in Germany – the CDU:

Politically it is just a fairly awkward attempt at siphoning support for Fridays for Future and climate change concerns for their own conservative agenda. But the more interesting part is the background image which in several aspects is in blatant conflict with the physical reality.

If you look at the image – use the link to see a larger version – you can find a number of typical beginners’ mistakes when doing whole Earth or large area geo-visualization:

  • The atmosphere is completely unrealistic in thickness and density profile. The actual earth atmosphere is relatively thick (meaning optically thick) in its lower part and thins out rapidly in the upper parts. The thick lower part at the full size image would be no more than 3-4 pixels in size at the edge of the earth disk – it would not be the fluffy kind of halo around it as shown in the above image.
  • The mountains are so excessively exaggerated in height that the whole thing becomes a caricature of the actual earth shape. This means the whole illustration completely lacks a sense of scale for Earth and the viewer gets the impression of Earth being a kind of toy model of maybe a few dozen kilometers in size.
  • The lighting is absolutely ridiculous – sun direction is kind of an early summer morning situation – though probably from a bit further north than physically possible. But this is combined with a nighttime visualization of the earth surface on the right (eastern) side – where the sun seems to come from.

So overall what does it communicate? That the CDU is making politics for an imagined toy planet that renounces the laws of physics.

September 12, 2019
by chris
0 comments

Clearer mountain views

I have been working during the past few months on a number of technical improvements for better quality in 3d views. These i want to introduce here together with a number of images from the Kashmir/Karakoram/Pamir region which is well suitable to demonstrate these enhancements.

Removing shading

All satellite images as taken feature the specific illumination due to the position of the sun at the time the image is taken. If you visualize the satellite image directly the shading due to the directed lighting is essential for reading the image. The flexibility in choosing a specific illumination in most cases is limited to selecting between different times of the year although in certain situations you have more freedom as i have written on several occasions here in the past.

Most people who produce 3d visualizations based on satellite images use this inherent illumination as recorded as the basis of their visualizations. This leads to a complete lack of flexibility with regards to illumination and also to subtle inconsistencies because the shading visible does usually not accurately match the 3d geometry shown and even subtle differences here lead to a lack of realism being perceived by the viewer and distracts from the actual content of the visualization.

Therefore my 3d views feature a precisely calculated simulation of the illumination individually selected for the view that is independent of the illumination when the image data used was recorded. Many of my views for example show an evening lighting although the satellite images used are recorded in the morning. For this to work well i of course have to remove the shading effect from the image data. I have been doing that for more than ten years now. The process for this has been refined significantly over the years and the most recent improvements in particular led to better accuracy and more robust dealing with difficult situations. Here an example.

 

Original L1C image

Atmosphere and shading compensated

As you can see the shading compensated version looks annoyingly flat and structure-less – but this is exactly what it is meant to look since the actual impression of the earth surface topography is to come from the specific simulated shading calculated on top of this in the 3d rendering. And while you can’t see the relief structure any more looking at the image the actual differences in surface coloring are better visible after the shading is removed.

Clearer geometry

The other improvement i have introduced here is a new processing of the geometry data to reduce noise while preserving acuity. I made use of the ALOS AW3D30 relief data – which offers fairly good coverage in the region in question. This data – like all other similar data sets – features a significant level of uncorrelated noise which is well visible when you use it in rendering directly. Reducing this noise while preserving the actual relief features leads to a clearer and better readable rendering. The technique used is related to the methods i use for producing generalized shaded relief rendering in 2d.

Original elevation data with noise


With noise reduced and generalized relief data

Examples

Here are various examples from the larger Kashmir/Karakoram/Pamir region rendered using the techniques described. All of these and more can be found in the catalog.

First for comparison the iconic K2 view i show on the main page on services.imagico.de in its old and new version.

K2 view from 2006


K2 view from 2019

And here a selection of further images of this region.






August 16, 2019
by chris
3 Comments

SotM 2019 Heidelberg program remarks

In a bit more than a month the State of the Map conference 2019 in Heidelberg will start and since the program is available now – including abstracts – here a bit of analysis of what we can look forward to.

I looked over the talks and workshops in there and made a rough categorization. This is not meant to be exact science of course – there are obviously borderline cases. Lightning talks are not included in this analysis.

  • There are about 25 program items from commercial organizations of some kind. Roughly half of these are big, international corporations. Most corporation have only one talk in the program.
  • There are about 24 program items from non-commercial organizations of some kind – from universities and public institutions to non-profits and organized community projects. About five of these are from the OSMF (who runs the conference). Apart from the OSMF no single organization seems to account for more than two talks.
  • There are about 22 program items i would classify as from individual community members or non-organized community projects.

As usual without being able to look at what spectrum of submissions these were chosen from it is hard to say much about this selection but it seems clear that care was taken to achieve a balance of different types of talks on many different levels. The most obvious bias is for English language – which i would be very much in favor of relaxing as a requirement. But that is just my opinion. Further thoughts on the general concept of SotM you can find in my post from last year.

Recommended talks

From a quick look over the program here a few recommendations on what to me seem to be very promising program points. This is not meant to dismiss any of the other talks as less valuable or interesting – it should be understood as a list of program points that stand out in my eyes, maybe as a suggestion what to visit if you can only see a handful of talks and have no more specific thematic interests.

  • Introduction to OSM: How it’s made and how it’s used – while probably not offering a lot of new insights for experienced visitors the concept of this program item promises to be fun and certainly worth visiting.
  • Communication and Knowledge Transfer in OSM – although i fear during a 20 minutes time slot this will hardly do more than quickly touch the subject this is a topic of high importance and central to the future success of the project. Hanna will probably provide both a solid summary of the status quo and a valuable perspective on where the deficits and problems are that need to be worked on.
  • Mapathon, mapathon, mapathon! – the title could certainly be improved, the main topic here is a critical look at established practices in collective community mapping endeavours (a.k.a. mapathons). Séverin and Nicolas have extensive experience in cross cultural cooperation on mapping in OpenStreetMap and the talk promises to bring this cross cultural dialogue into the conference with participants from different countries contributing their thoughts and experiences directly. This talk is the starting point to a number of program items following up on similar subjects – a second talk directly afterwards and a bilingual discussion session later.
  • Is the OSM data model creaking? – critical reflections on the fundamentals of the data model are rare both in OpenStreetMap and in the GIS world. This talk promises something along these lines – a subject of high importance in my eyes. Since the focus of the talk is on a fairly specific use case (cycling maps and routing) there is a chance that it will end up talking about relatively cheap workarounds specific to this use case but it might also take a more fundamental look which would be particularly interesting.
  • New processes to agree on tagging suggestions and their interaction with the editing software available on openstreetmap.org – this might serve as a more specific followup to the Communication and Knowledge Transfer talk mentioned above. Roland always provides a sober analytic perspective on things. He also takes the innovative approach here to specifically invite contributions from people who are not at the conference. This way the session seems designed more as a followup to a process of collecting ideas on the subjects via digital channels. Independent of the subject – which is of high interest in itself – this is an interesting and innovative approach for having a discussion, developing and exchanging ideas and connecting remote digital communication with an in-person workshop.
  • OSM Vector Tiles in custom coordinate systems – Coordinate system agnostic map rendering is a matter almost completely ignored by the cartographic mainstream in OpenStreetMap. Like with the OSM data model talk this might also turn out to be relatively superficial but there are so many interesting problems related to this topic most of which most map producers are unaware of that this could very likely be interesting for people producing or who want to produce maps in other projections than Mercator.

Scholarships

One other very interesting program item is Scholar Lightning Talks which provides the first publicly available piece of information on the otherwise extremely intransparent SotM scholarship program after the call for application. The list of talks does not necessarily include all OSMF scholars at the conference – there might be others who have a regular talk or give no talk at all. But it lists 10 scholars which is plausible to be the full set or at least nearly the full set of people that receives a scholarship.

So where do the scholars come from? We have:

  • Philippines (x2)
  • Nepal
  • Lesotho
  • India
  • Nicaragua
  • Netherlands
  • Gambia
  • Tanzania
  • Colombia

For comparison – last year we had:

  • Germany (x2)
  • Portugal
  • UK
  • Russia
  • Mexico
  • Colombia
  • Uganda
  • Kenya
  • Lesotho
  • Mozambique
  • Philippines
  • Nepal
  • Bangladesh
  • Indonesia
  • China

And 2017:

  • Bangladesh
  • Albania (x2)
  • Lithuania
  • Cyprus/USA
  • India
  • Argentina
  • Italy (x2)
  • Senegal
  • Niger
  • Nicaragua/Costa Rica
  • Sri Lanka
  • Sierra Leone
  • Colombia

Notice something? Well – first of all obviously SotM 2020 is most likely going to take place on the Philippines since the previous two SotMs took place in countries where the previous year there were two scholars from. For more patterns look at the following map showing the distribution of scholars from the years 2017, 2018 and 2019 (in blue the home of the scholars, magenta the conference locations).

Where OSMF SotM scholars 2017-2019 came from

As explained previously there is a clear geographic bias in where SotM scholars come from. This can well be illustrated in terms of statistical likelihood to become a scholar. Note this is an overall bias in the whole process – from being attracted and motivated to apply for a scholarship until actual selection. Due to the lack of any public information on where the applicants for scholarships come from (this has been kept secret by the OSMF despite inquiries) there is no way to determine if this bias primarily lies in the selection or already before.

  • If you are from Africa your chances to become a SotM scholar is much higher if you are from south of the Sahara (10 of 10 are from there) and significantly higher if you are from a former British colony (7 of 10).
  • If you are from the New World your chances are much higher if you are from the larger Central America region between Mexico and Colombia (6 of 7).
  • If you are from Asia your chances are much higher if you are from a former Britisch or US colony (10 of 12).
  • If you are from Northern Africa, the Middle East or Central Asia your chances are very small (none of a total of 41 scholars comes from one of these Regions).
  • Individual countries with a strong preference are: Colombia (3x), Philippines (3x), Albania (2x), Germany (2x), Italy (2x), Nicaragua (2x), Lesotho (2x), India (2x), Nepal (2x), Bangladesh (2x).

August 10, 2019
by chris
2 Comments

OpenStreetMap and collection of local geographic knowledge in 2034

Today we celebrate the 15th anniversary of the OpenStreetMap project. The date is somewhat arbitrary but i none the less want to take the opportunity to think a bit about what the next 15 years might bring and how things might look like when the 30th anniversary comes in 2034.

Most scenarios for the future of OpenStreetMap that have been presented in the past concentrate on practical aspects, how the project can scale as it grows, what effect technological developments have and what challenges this brings.

What i want to look at here instead is more the social dimension of the project as a whole. And for this i want to split what OpenStreetMap is today into two separate things:

  • The specific project with the name OpenStreetMap.
  • The general idea behind OpenStreetMap to collect and share local geographic knowledge people have of their own environment into an open data set through egalitarian self-determined cooperation of these people.

I can of course not reliably predict what will happen over the course of the next 15 years but looking at trends both in OSM and in the world in general over the past few years it seems likely that

  1. both these things will still exist in 15 years
  2. they will however separate and develop into different directions

What do i have in mind when i talk about separation here? The idea is that the OpenStreetMap project will probably move away from the focus on local geographic knowledge gathered and managed by local mappers. This has already been a trend over the past few years with the fraction of data in the OSM database that has never been substantially reviewed by a mapper with local knowledge massively increasing. Some might argue that we already had large volume imports much earlier but in contrast to early imports 5-10 years ago which were made at least with the intention that the data will be adopted by local mappers and integrated with their local knowledge – today, while there is still some superficial pretense in that direction communicated, this is not something that is widely believed any more when data imports and organized remote mapping projects take place. The fraction of data not grounded in human local knowledge varies a lot depending on where on earth you look and globally it might still be less than 30 percent but it is growing fast.

It is a pretty safe bet that this trend will continue for the coming years. What is less sure however is the future of cooperative collection of local geographic knowledge. We can only speak of separation if there is actually a notable community outside of OpenStreetMap engaging in such data collection. I don’t really see a likely practical scenario how this might happen so far. Many will probably consider this unlikely to happen at least as long as OpenStreetMap exists in some form because of the high hurdle in building a new community with OpenStreetMap as competition. Some might think of Wikipedia/Wikidata serving for this role but that seems unlikely considering the different fundamental premises of this project. I none the less think separation is likely to happen because the trends in OpenStreetMap described above will mean there will be much less room for the kind of egalitarian cooperation described. If the focus shifts away from local geographic knowledge gathered and managed by local mappers this will automatically lead to an increased social and cultural stratification within the project because there is no natural local ownership and control of the map data legitimized by local knowledge any more. And how do you motivate someone to manage and supervise mapping in a certain area and give them authority for doing that without this local ownership and control? By paying them and giving them social privileges is the obvious answer and we already had people calling for paid ‘community managers’ and we certainly will see initiatives in the OSMF in the coming years to hire or pay people for various management tasks. And even at the actual work level – like mapping and software development – we see an increasing social stratification fueled by some people being paid full time by external stakeholders to work on the project while the majority of local mappers forms kind of a working class with no substantial voice in decisions that matter – not even locally in their area.

This probably sounds fairly dark – it likely would not feel this way if things actually develop as predicted. Clear leadership within the project is something that is quite appealing to many participants. While i find the cross cultural egalitarian cooperation of people from all over the world openly sharing their local knowledge the most significant aspect of OpenStreetMap and the core of the success story of the past 15 years i know that many mappers don’t really care much about this and essentially just want a smoothly running platform where they can contribute stuff and do mapping work they enjoy. And there is nothing inherently less valuable and justified in these preferences than in mine.

As said i am quite unsure how the future of cooperative collection of local geographic knowledge will look like. It is quite possible that this will take the form of smaller regional projects, potentially with a strictly limited thematic scope. Right now any initiatives in that direction are essentially absorbed by OpenStreetMap in most cases and it is hard to predict what would happen if this was no more the case. It is also possible that the time scale of such developments exceeds the 15 year horizon i have looked at here.

Is there a possibility that this separation does not happen and OpenStreetMap is still or even more than today and in the past 15 years be a place for collecting local geographic knowledge through egalitarian, self-determined cooperation of individuals in 15 years? Of course. It would require a lot of commitment from local mappers all over the world to defend this vision of cooperation – even against strong interests incompatible to that. I would like to see that happen but as explained trends at the moment point into a different direction.

August 2, 2019
by chris
0 comments

Europe views based on Landsat color data

Some time ago i introduced a new technique for producing accurate color mosaics from Landsat and Sentinel-2 data using pixel statistics methods. Here now some larger area 3d views of Europe produced using this imagery allowing for a more accurate and more balanced color reproduction than previous renderings of the same area produced using MODIS data.

   

July 30, 2019
by chris
0 comments

Green Marble based globe views

I have a bit of a backlog of 3d views i have not yet introduced here on the blog and which i am trying to work through in the next few posts.

The first is a set of globe views based on the Green Marble mosaic. Whole Earth views possibly seem at first glance a relatively boring variant of 3d Earth views because of the lack of variability in composition. But the options in perspective are more diverse than you might think at first.

Here a sequence of views with the same lighting centered on the same point but with different distances from the Earth center and to create matching sizes of the planet in view different camera angles. As you can see although the appearance of the planet as a whole is always circular the overall impression varies quite a lot.

400000 km distance

40000 km distance

20000 km distance

10000 km distance

This is the combined effect of the inherent distortion of a wide angle camera at shorter distance and the curved shape of the Earth surface. It allows for significant variability in composition of a whole earth view even when looking at the same point on the earth surface.

Here the set of views i want to introduce – which i had produced quite some time ago but which i added to the catalog just now:


     

Of course custom views are possible as well – just drop me a note if you are in need of one.

July 15, 2019
by chris
0 comments

The Arctic is burning – or not

Someone at some point in the last few weeks seemed to have managed to put the idea into the mind of journalists that there are wildfires burning in the Arctic and this being a exceptionally newsworthy story somehow. These reports are so full of lack of background knowledge and made up nonsense and exaggeration that they are a feeding frenzy for all the climate change deniers out there so i thought it would be a good idea to clear things up a bit. Here a few factual points on the matter.

The Arctic is commonly defined as the area north of the climatic tree line. Roughly that is the area from some short distance south of the Arctic ocean coast northwards in Siberia, Alaska and Canada plus Greenland, the Bering Strait, the Aleutian Islands and Hudson Bay. You can see this approximately illustrated here. The Arctic Circle is sometimes considered an approximate geometric limit for the Arctic with the actual tree growth limit being somewhat further south in parts of North America and somewhat further north in Siberia. There are only few wildfires burning at this time of year (June to early July) – both this year and in previous years – in the Arctic. In June to early July snow melt is either in progress or has just finished in most of the Arctic, the ground is still wet even at the surface therefore and fires are rare.

Forest Fires Siberia 2019-07-04

The fires the current fuzz is about are forest fires in the boreal forest. Such forest fires happen every year in significant numbers and are in principle a natural occurrence. This year they have started relatively early due to widespread warm and dry weather in June in particular in Alaska and parts of Siberia.

Forest Fires Canada 2018-08-15

Forest Fires Siberia 2018-07-03

But generally speaking there is no clear indication that there is something exceptional about the boreal forest fires this year in their overall extent compared to previous years. In particular last year there were very extensive forest fires in Canada and we also had significant fire activity in early July already. Ultimately this can of course only be assessed after the end of summer.

If there is a general increase in forest fire activity during the last decades due to climate change is a question that surely deserves a closer look. But this is nothing you can determine from observations of one or two years. Also trying to count individual fires through remote sensing is not a suitable instrument here. And if there is a long term increase the real question regarding climate would ultimately be if and how this affects the carbon balance of the boreal forest as a whole. That is not a trivial question.

The most likely explanation for the fuzz in the media is that there is a larger number of researchers trying to push sensationalist messages about extraordinary observations to generate publicity for their work. And they find willing allies in journalists desperately looking for clickbait material. This is not very responsible. The lack of thoroughness in analysis of the observations and leaving out the broader context plays directly into the hands of those who want to discredit climate research as flawed and non-credible.

July 1, 2019
by chris
0 comments

Summer and Winter 2019

Some satellite image impressions from mid 2019 – all based on Sentinel-2 data.

Cloud vortices around Heard Island near mid winter:

Last sea and lake ice thawing near the coast in northern Canada:

Another southern hemisphere winter impression from Kerguelen:

And some beautiful algae bloom around St Kilda off the Scotland coast:

All images are available in large on services.imagico.de.

June 19, 2019
by chris
1 Comment

On surveying the Surveyors

The OpenStreetMap Foundation board has published a summary of the survey they did leading up to their face-to-face meeting in May. This contains a lot of interesting insights but also a number of rather questionable ideas and tendencies. Here i am gathering a few of my thoughts on the subject of surveying the views of the OSM community and the larger field of communication and social dynamics in the project.

As background – this relates to things i have written in the past on this blog – like here, here, here, here and here as well as recent comments on the survey and related matters.

The diversity of OpenStreetMap

Quite a lot of people view OpenStreetMap as yet another garden variety crowd sourced internet project – with the main difference to most of the thousands of other similar projects that exist being that it has become significantly larger than most of them. This is a misconception though.

What makes OpenStreetMap stand out compared to other internet communities is not its size but its diversity. And with diversity i don’t mean the individual lifestyle choice and personal freedom of expression diversity but fundamental cultural diversity. While mapping in OpenStreetMap is dominated in numbers by people from urban Western Europe and North America mappers there are strong local communities in particular in Russia, Latin America and Japan – communities with their own cultural identity who are not just integrating into and adjusting to the cultural values of the Europeans and North Americans but who locally and independently define their own cultural style of mapping while still seamlessly integrating into the global project.

This is what makes OpenStreetMap special and unique compared to other internet communities. And the amazing thing about this is that it happened without this being planned and engineered by some central authority. The idea to record and share your knowledge of the local geography where you live with others through an open database is something that resonates in many cultures all over the world.

This is very important to keep in mind when talking about communication and social mechanisms in the OSM community to not fall into the trap of taking cues from other projects on how to manage things and trying to shape OSM in their image without critical reflection on if such ideas are actually suitable for OSM.

The survey

Due to the cultural diversity of the OSM community surveying the views of this community in a representative form is practically impossible. The survey the OSMF board made does not actually try to do this but despite the fact that this is acknowledged when interpreting the results certain ideas about using the concept of surveys as a political instrument shine through that are highly questionable.

With the survey being non-representative and therefore without meaningful quantitative results it inevitably produces contradicting answers. In this case there seem to have been for example answers calling for making imports easier as well as calls for a more meaningful regulation of imports. Without further information this can be used to justify just about any political initiative w.r.t. imports in OSM.

The best way to regard this kind of survey results (at least in the summarized, aggregated form they are presented here) is to consider them to be the results of a highly selective brainstorming session for subjects that might be of interest. If this helps the board members recognize matters they were previously unaware of that is very useful. But using the results to pat yourself on the back because you can see the subjects you find important are also considered important by at least some people who have participated in the survey is not a good idea. This you would not need a survey for.

To me the survey results as presented are not really that astonishing. They more or less confirm the image i had of the views and priorities of the more articulated parts of the English speaking parts of the international OSM community (and with English speaking i here include also those who speak English regularly as a foreign language). The only thing i kind of missed is the organized editing matter (which might be considered part of the commercial influence topic). This might be however due to prioritization since the survey specifically only asked for the most important topics.

Surveying as a substitute for open discourse

The more problematic aspect about the board’s report on the survey however is the idea of making community surveys a regular part of the OSMF board’s communication with the OSM community. I am quoting here the beginning of the third paragraph – which is fairly difficult to read but i would urge everyone to read this carefully because it ultimately is very explicit in outlining a certain idea for how communication between the OSM community and the OSMF board is to happen:

OpenStreetMap is made up of everyone who takes part in it, and hearing each other’s voices in a coordinated way, on a regular basis, will help prioritize where work is needed and what actions to pursue. It ensures a standard way to engage. Surveying can help set up mechanisms to route issues to the right place, share pathways for OpenStreetMap members to contribute and address problems, and identify where the project as a whole needs more help to come up with answers.

This could have been directly copied from a corporate handbook on PR or human resources management – where surveys are immensely popular as well since the institutional hierarchy is valued and direct communication between higher management and customers or individual employees is usually frowned upon.

What the OSMF board seems to say with the paragraph cited above (though i am not really sure if all the board members were aware of what they are actually saying here – it is fairly convoluted language) and what also got reiterated by a board member in the June 19 board meeting in similar form is that they want to adopt a similar principle in the OSM community where the input of the individual community members to the management is to happen “in a coordinated way” and not through open discourse.

What you have to keep in mind (and what is also frequently forgotten in the corporate world sometimes leading to drastic failures in management) is that a survey always exclusively transports wants and desires, it can by design never serve as an instrument of arguments and reasoning. Reducing communication to a survey cuts you off from arguments and reasoning and prevents you from understanding the underlying motives for people’s desires. Reducing people to their basic wants and desires is also ultimately quite demeaning.

The role of the OSMF board

The motivation for this kind of approach is quite clear. In a growing OSM community it is becoming increasingly difficult to identify in the cacophony of opinions articulated where there are actually real needs for things to work on. But in my opinion to a large part this problem arises from the OSMF board engaging on matters that should not be their concern. The task of the OSMF board is not to govern OpenStreetMap, it is to provide a framework in which the community can govern itself. And the design of this framework should not be based on the momentaneous wants and desires of a non-representative part of the OSM community as gauged by a survey but should derive from the long term values of the project (which i have written about frequently in the past – see the links above).

Now i don’t want to sweepingly reject the idea of doing surveys of the OSM community. As a supplement and to provide input to an open discourse these can be of value. And with published free form answers surveys designed and initiated from the community could even become an integral part of an open discussion. But ultimately discourse on matters in an open and diverse community like OpenStreetMap will happen across many different channels in many different languages and styles of expression – inevitably inconvenient for any centralized management attempts. Any attempt to channelize and control such discourse from a central authority is bound to fail in one form or the other.