How OpenStreetMap can support the UN Sustainable Development Goals (SDGs)
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 70 | |
| Author | ||
| License | CC Attribution 3.0 Unported: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/58516 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
State of the Map US 201924 / 70
5
7
14
20
26
28
30
31
33
36
38
40
43
44
51
53
55
56
60
61
67
68
69
00:00
Lecture/Conference
01:36
Computer animation
07:28
Computer animation
13:21
Computer animation
13:46
Computer animation
19:39
Computer animation
21:35
Computer animationDiagram
Transcript: English(auto-generated)
00:00
Hey everybody, my name is Ross Burnett. I'm excited to present today Excited to be the first presentation and This is my first state of the map, but it's great to see a bunch of familiar faces in the crowd It's I've heard nothing but great things about this conference. So I'm looking forward to it I'm gonna be talking about how OpenStreetMap can support the monitoring of the UN sustainable development goals
00:24
So who am I? I'm Ross Burnett. I'm a project manager at Azavia Azavia is a software company based in Philadelphia We prioritize working with open source working in the open and We're a B corporation with a mission to advance the state-of-the-art in geospatial technology and apply it for civic social and environmental impact
00:44
So the sustainable development goals are completely aligned with the type of work we want to do It's been relatively new. We've been working on the sustainable development goals for about a year now I went to the first conference to kind of explore that world last year
01:00
in Dubai at the UN World Data Forum and just like it's a whole There's so much different efforts and work going into it. It's a pretty exciting space So my talk today Intro, that's what's happening right now. I'm gonna give a broad overview of the SDGs So just to help provide context for the depth that I go in
01:23
Could you please raise your hand if you're familiar with the SDGs? Cool Looks like about half of you. So I just have a few slides that are gonna talk about it a little more in-depth Then I'm gonna do a deeper dive into one SDG indicator in particular. That's gonna be 9.1.1
01:40
Azavia has a project working on evaluating indicator 9.1.1 for the entire world Discuss a little bit of the technical approach of doing that Talk about opportunities challenges and next steps and then leave a few minutes for questions at the end Azavia has been working with OSM for the last couple years and
02:02
It's still there's it's such a rich deep community and I'm curious because I'm certain there's folks in the audience that are Doing related work, and I'm really interested in hearing about your challenges also So I'm hoping to start some conversations either in the room or beyond the beyond the talk
02:20
So broadly speaking The SDGs there's 17 they are global it's a global framework for Measuring for helping to measure and set targets for the agenda 2030. These are where do we want to be? as a global community of humans in 2030, these are very ambitious goals like
02:43
Ending poverty ending hunger ending climate change these very overarching like hard to argue with these are good goals And But to help measure the target to these goals Each goal is broken down in between about one and five targets And then the targets are broken down in between one and five indicators and these indicators are quantitative measures that can be
03:07
tracked and Compared over time so we can see how how is our progress towards? toward towards the goals There's so much Open data that can be used to support the SDGs and a lot of open geo data. I gave a talk similar to this one
03:26
In April at phosphor G in San Diego With a little more focus on satellite imagery but I find a lot of overlap between the type of things that are observable with satellite imagery and what in terms of measuring monitoring observing
03:41
Geospatial data and what's available with open street map so geo the group on earth observation put out a report a couple years ago where they evaluated all 17 of the indicate of the SDGs and then mapped out which ones have either a strong or like totally geospatial component so some of the indicators like
04:03
increasing the reading level of the average Average youth is not exactly geospatial although many of the indicators Are discuss a proportion of a population these start to be questions that are geospatial and then some are extremely geospatial in nature in terms of
04:22
What percentage of your country is covered by forest and so and I'm going to discuss a few more that are? Very are more closely aligned with open street map so came across my attention That Tyler Radford gave a talk somewhat similar to this one a few weeks ago in Stockholm
04:41
At the Wikimania conference, and it was exciting to see how he's how hot on the humanitarian open stream app team Is thinking about some of these same same types of questions? He went into a lot of more in the field Data collection he gave a few examples of working in Tanzania
05:02
Working with folks to collect points of safe drinking water and helping measure all the sewage Open sewage that was in Dar es Salaam and he links to a report that I copy at the end and the resources on a number of tools that they found I actually reached out to Tyler a couple days ago and
05:21
Asked if I could include this slide. It's sort of a mash-up in between in between a couple slides And he's like yeah, we got a we got to figure out how to collaborate. This is this is great So it's for me part of the reason that SDGs are exciting is because there's this it's like it's a framework it's a way of talking about a lot of these same goals that we have but
05:43
People have spent a lot of time thinking about these metrics and how we can measure them and so having this kind of common Framework we're talking about it. I find I find really helpful so going into a few of the specifics of I'm gonna go through about five indicators That you can clearly see or like if this information was in OpenStreetMap
06:02
We'd be a pretty good way into reporting on this indicator. So the first one is 11.7.1 And so the way these work is this is goal 11 Target 7 indicator number one, so there'll be a few different few different targets for each goal and then a few different indicators for each target and
06:23
So this is the average share of the built-up area of cities that is open that is open space so we can we already know That there's a lot of parks and open governmental land and kind of open land And if we knew that it was just perfectly mapped we could take that data from OpenStreetMap and then directly have this
06:41
Have this indicator calculated Another one is the proportion of the population using safely managed sanitation services So in developed countries we take for granted a lot that we have access to clean drinking water This is clearly not the case for much of the world And there are you could imagine adding into OpenStreetMap. Here's here's a hand-washing facility. Here's X. Here's a
07:05
Faucet here's here's water and so having in a sense of Coverage of these different features could be really helpful for helping to understand how we're doing in terms of clean water and sanitation Similarly we take for granted access to ATMs and banks not the case for much of the world
07:23
How are these services distributed where there are gaps if that information was in OpenStreetMap You could start to have a better sense of tracking tracking these indicators Proportion of important sites for terrestrial and freshwater biodiversity again
07:40
Coverage of protected lands that information we could imagine being being in an OpenStreetMap to help understand how these areas are protected And the one that I'm going to go much deeper in is SDG 9.1.1 This is a proportion of the rural population who live within two kilometers of an all-season road
08:03
I am excited about this one in many ways because it is relatively straightforward In principle you can understand we're looking for roads, and we're looking for the population within them Then I'll go into it in a little bit in terms of well What exactly is all season and what exactly is rural?
08:22
These are questions that you need to start really getting into the weeds in but compared to some of these other indicators It's there's the data sets that you need are clear So I want to take a second to walk through the kind of process of what is this the system of SDG Of the targets and indicators so we kind of have the shorthand version of the title industry innovation and infrastructure
08:46
So what what does that mean? That's just a shorthand to kind of fit onto this nice pretty UN format So it's kind of compelling image that you see all over the place once you enter this SDG world But the actual title of the goal is a little bit longer
09:02
building resilient infrastructure promote inclusive and sustainable industrialization and fostering of innovation You're like, okay, that's a little more meaningful, but still don't know exactly what that means or how we could measure it Then the target attempts to clarify it a little more specifically We're developing infrastructure
09:20
The purpose of we're doing this is economic development human well-being and we want it to be affordable and equitable So we want we want to make sure those rural populations are being serviced in addition to our urban populations and then the actual indicator this is We can put a number to this we could calculate this for a country and then at the end have a proportion
09:43
70% 80% 100% and I've had this conversation because I think it's important to be critical of these. It's like is building roads actually good We kind of take that we kind of take that for granted But there's actually cases where that might not be good roads bring in or it's can bring in crime roads can lead to deforestation
10:03
Roads can bring all kinds of negative consequences with them. So it's this constant balance and being critical but The point of the goal is getting that access to services like health services economic opportunities Education and so roads are being used as a proxy for access to these these different
10:24
Services that that we do more black and white assume are good So a lot of the indicators have this sort of we have to you have to evaluate and are there ways to take that into account So, how are the SDGs being monitored now No one organization has access to all the data that we're gonna need
10:44
To calculate the indicator value for any any of the SDGs We couldn't even possibly hope that it would just all magically be an OSM But it's this concept of piecing together the different data sets Knowing where they're at So I'm going to talk about how how they're being measured because it's it's a massive system. There's over 230 indicators
11:06
And 169 targets and some of them are much more Clearly defined and so we have the system of the UN created the system of tiers. So each of the indicators is Assigned a tier so a tier 3 means the methodology is not even established
11:22
Even in the example that I'm gonna walk through a little bit more nine dot one dot one It's relatively straightforward, but you still get questions. Like I'd mentioned what's rural What is all season? Should two kilometers should we just stick with two kilometers or is two kilometers the same in a mountainous region as a
11:42
Flat region are we if you live across the river from a road? Do you still have access so you get into this? nitty-gritty of Like disagreement it's like well that indicator doesn't make sense in in my area. It doesn't make sense in my city And so Establishing that methodology there's working groups. There's calls There's a lot of information and in meetings and discussion that go into this
12:03
So advancing from tier 3 to tier 2 is actually fairly significant and complicated process So once you hit tier 2 that means the methodology is clear But the data is not regularly being collected in the in nine dot one dot one. We we have access to that data
12:22
But Some some of at least some versions of that data but for many of the indicators how how are we gonna measure the reading level of the youth do we have a good way of Capturing that information so you can't get to a tier one unless that data is regularly being produced For 50% of the countries and you have a clear established methodology. So this tiered system helps understand
12:44
priorities for funding how are some of some of the SDGs are like Receive lots of funding some receive a lot less Is that is that being equitable to kind of the needs and in the ability to to measure and manage? So in addition to the tiered system you have this concept of a custodian agency each of the 230 or so indicators is assigned
13:07
UN custodian agency So these are groups like the World Bank UN habitat World Health Organization and they work with the countries They work with the statistic offices To refine that methodology and to report on it
13:20
and I was grateful to see that some of the Some of the custodian agencies are really open and excited actually about open street map So the World Bank who's the custodian agency for nine dot one dot one? there's a He wrote this recently in a blog, but they are actively promoting the use of open street map
13:41
and introducing it to two countries to Just start to include that in the official statistics So Then there's this concept of the National Statistic Office Most countries have one some countries have like it's more distributed. The u.s. Actually does not have one NSO We're divided between all sorts of different groups
14:00
but a lot of countries have one National Statistic Office that is responsible for reporting on things like GDP and trade and generating official statistics for a country in some cases The NSO is just burdened suddenly like do everything you're doing and do SDGs on top of that Some cases you have countries where there's someone dedicated to measuring and monitoring reporting on SDGs
14:23
So you see quite a wide variety and we could imagine some countries are more equipped to gather and collect this data some don't have capacity and So understanding that landscape is helpful to understand the state of state of the SDGs So, how did Azavia get involved with this work?
14:41
Azavia created an open source tool called raster foundry, which is a tool for using satellite imagery created in partnership with NASA and Department of Energy used to collect and make available open data sets and A group at the UK Office of National Statistics saw raster foundry wanted to include it in something called the UN Global Platform
15:02
This is a tool that is to be distributed to National Statistic Offices to help them calculate The SDGs amongst other things So we started working with them and then we said okay, what's one indicator that potentially we could calculate within raster foundry Or within the the UN Global Platform and sort of this is happening
15:25
Maybe about 12 maybe about last year 12 months ago or so And so we're at a conference at the World Bank and a few folks at Azavia met a group at International development organization called Cardno and they had a project running with DFID that's the Department for International Development at the UK that was
15:45
tasked with helping countries in Asia and Africa actually build roads But a part of their funding was dedicated towards monitoring the progress and access of folks to the roads and they had been calling a rural access indicator that's essentially synonymous with
16:02
Indicator 9.1.1 and we said hey, we're looking to Develop the capacity to do something at a global scale potentially with open data They're working they work closely in country with departments of transportation various NSOs to actually collect collect the data that you need and we
16:22
Developed this kind of pilot project. That's actually started about a month ago And so we have some preliminary results at all. I'll share in a little bit So we said, okay, how do we how do we actually measure this thing globally? Again, 9.1.1. We just need three data sets in principle access to roads
16:41
access to countries population information and then these rural boundaries and Thankfully, there's open data Globally that's available for each of the three indicators. We have open stream out for roads We have the NASA gridded population the world data set. That's a raster data set I think it's 30 meter or 100 meter resolution of
17:02
population and then grump the global rural urban mapping project which defines urban and rural boundaries and so each of these data sets is incomplete But we chose them as a means to start kind of this benchmark. Can we can we get some initial values?
17:25
And So the reason I'm excited about it is yeah, it's it's not going to be the most accurate Especially given all the considerations of open street maps not equally mapped in in all countries and Maybe grump defines urban and rural that doesn't apply their methodology doesn't apply across countries
17:44
So it's important to work within countries and talk to them ideally get a country's data set And to be able to make that comparison so Part of my premise and hopefully we'll have our final products Supposed to be in the next month or so
18:01
But the concept is if we go present at a conference and say hey country. Hey. Hey, Colombia your Stg 9.1.1 value is 60% and they're like no we think it's higher than that and we're sort of hoping to encourage a Conversation we want to make lots of caveats like this We don't think this is the final number but using the best available open data using this methodology
18:23
Here's what we arrived at If you have better data ideally let's get it into open street map But otherwise let's compare it and see see how it's different and everything we're doing is open source transparent and not trying to Yeah, not trying to go be we're trying to work with the countries
18:43
So some of the challenges we're facing our definitions data quality and methodology This is in terms of yeah, I've mentioned it several times now, but what's real what's urban Is this data good enough and then? Methodology both in terms of how you're actually generating that number, but then is it possible to scale this up?
19:05
So that was part of the opportunity that was interesting to Azavia Can we do this at this global scale and what are the implications of? Using all of open street map in doing a global calculation So Something in addition, so we have this all season. So what makes something all season the concept is we don't want
19:26
We don't want a dirt road that gets flooded out Half of the year to count as actually access to these different services. So can we rely on tags? there's a tag in open street map that's surface type and So one of the preliminary maps we have we did a demo for South America as kind of a prototype and
19:46
In yellow, we have all the tags of roads that are like we could consider paved and I'll go through What that might mean in a second? Green teal is like pretty clearly unpaved and then you can see the vast majority of roads are undefined
20:00
There's no there's no tag for the surface type And so for South America, there was about seven million road segments And the vast majority Were no just nothing put in there Lord knows when I'm adding roads to open street map. I'm mostly just adding roads. I'm not worrying about all the different
20:21
All the different features for better for worse And so the top handful we can recognize we can say okay, I kind of know what that means But I created this very fancy data visualization tool yesterday to show there's over 400 400 tags and you go down like past the first 10 and you start seeing characters that you don't recognize and
20:45
salt and and you and all sorts of things that like inconsistencies and so these It was like two-thirds of the tags had only one or two instance type Instances of that and so if we're gonna rely on open street map for this
21:01
Clearly we got to figure out a way to take this take this into consideration So how did we do the actual analysis for this at a global scale The team that I work on Creates an open source tool called geo trellis. This is an open source Scala library for primarily working with raster data But increasingly we've been working
21:21
With open street map and vector tiles It leverages Apache spark to do distribution and then another more recent library at Azavia called vector pipe for working with open street map data and then visualizing the results with Vectortiles, so what we do is we clip all the geometries to a regular grid
21:43
And then find the center points we project a UTM so we have a standard meaningful Two kilometer buffer if you just had the web Mercator that two kilometers very quickly doesn't make any sense And then do a summary with the NASA population data
22:04
So if you want this demo for South America is available online demos.azavia.com, but rather than try and do it live I just took a little screenshot so more or less. We've collected the values for each country And then just to help understand the different
22:20
The different distribution of the streets added those in as vector tiles and then the blue dots are map box Data visualization tool for showing hospitals, so I just pulled the forget it was 10,000 or so Anything labeled as hospital from open street map just to start to get a sense of how how are these distributed and how what's what's the access?
22:41
Like so this is sort of a preliminary version, but starts to get a sense of how How are these things distributed? So some of the lessons learned and these are actually from some of the notes with our partners over at Cardno But I think there's gonna be some talks later about different road agencies and how they're working with open street map But our partners are constantly sending emails though, we just met with the group in Ghana
23:04
We just met with the transportation transportation agency in Tanzania there. There's several of them. They're not talking to each other They haven't heard of open street map And so just there's like an education component of is this And as far as they're concerned our partners are big advocates of open street map and they're like it's the only integrated
23:24
Platform that they're aware of where the different road agencies can work together And they are saying that if the government agencies are involved in the OSM update Then the national statistic offices are more likely to endorse endorse that data is official There's a couple countries that it's like this sort of semi official status if that's even possible
23:44
but For these SDGs you need this like authenticated This is an official recognized data set and each country has different processes for what that looks like So what's next some open questions that I have Is global public data good enough to measure the SDGs? I think for me at least it's pretty compelling for 9.1.1
24:07
I'm certain the data quality for others is like not nearly as good in large part We picked 9.1.1 Because you could have a sense of like it's it's fairly complete at least in some countries in some regions So we can have a sense of sense of it But can open street map be good enough for other SDGs and then what is this process of authenticated and validate validated data?
24:27
How do we have consistency and tagging types and this country endorses this data and is monitoring it? And then what role do private companies have in SDG monitoring? There's different funding sources. Do you work directly with countries? Do you work with philanthropists or groups like the World Bank?
24:47
And then some of the conversations that I'm hoping to have in the in the next couple days are Just is work that you're doing already aligned with an SDG There's like they're pretty dang broad. There's a document that I'll link to and I can share this out
25:01
I don't know if there's an official means or I can tweet it out the presentation, but there's a link to all 230 Indicators and it's pretty easy. I sometimes if I have a new interest in mind I'll like scan through that list and think like okay, this could be relevant or this is not relevant But again, I think the SDGs are useful framework for kind of funneling a conversation
25:22
I want and I'm certainly not like they're they're just the best and they're gonna solve all their problems It's like by 2030 everyone kind of recognizes. We're probably not gonna get there But at least having a focus for the conversation contributing to OSM always good And then sharing. Yeah, are there other data sets either?
25:45
Either already in OSM or could be imported into OSM that might help monitor at the SDGs So here are some resources for the That geo publication. I mentioned here's the group from hot OSM or the the guide from hot OSM that talks about OSM plus
26:03
the various SDGs they're looking at The full list of indicators and then the demos So I think I'm right on time. That's what I got. Thank you very much
26:20
I should have said does anyone have questions? Yep In terms of the OSM
27:00
Yeah, yeah, thanks Matt. That's a great question. For my partners working with Cardno They said just a lot of these road agencies in a lot of countries don't don't know about OSM and are maybe more suspicious of OSM and I think there's a there's an education component of like actually this this is better data than
27:24
Then we see in a lot of cases so I think part of it is just like Continuing to talk about and sort of At what point and I'm sure there'll be other talks about this But at what point is this good enough to be an official data set? We could imagine there's gaps in even the official data set. So I think
27:42
If you have any ability to talk with road agencies and in Africa, that's that's a great option And then I think I'd be interested to see other examples like this. I think doing it Globally for a lot of the data sets is probably unrealistic We chose this one because I think it's one of the lowest hanging fruits
28:00
But maybe you could do it at a city scale or maybe you could do it At a national scale of like let's coordinate that one that was like mapping all the parks Can we get a reasonable estimate on that number by like having a concerted effort to map the park? So I think I think that's probably one way. I think the tools are available. It's a matter of coordination and is this is this important?
28:23
Yeah Yeah, actually I forgot to mention but my colleague Eugene Chapish will be given a presentation focused on vector pipe and comparing
28:41
OpenStreetMap buildings to the Bing building footprint data set that was released and we did a national comparison of Just a building matching algorithm So roads are definitely more difficult to do at that type of scale and we haven't we haven't really got there yet But we want it for in this case compare
29:01
The global data that's available with the in-country specific data So that's I think it's just getting started and definitely that's gonna be essential. Yeah
29:31
Yeah, I
29:42
Haven't heard about anything about that So I can't speak to that unfortunately, but yeah, I'd love to chat more and on that note. Thanks everybody I'd love to have conversations with folks about this in the next couple days
Recommendations
Series of 16 media