Data is USELESS!
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 | 295 | |
| Author | ||
| Contributors | ||
| License | CC Attribution 3.0 Germany: 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 | 10.5446/43421 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
| |
| Keywords |
00:00
9K33 OsaSlide ruleDifferent (Kate Ryan album)Observational studyComputer animationLecture/Conference
00:37
Geometry9K33 OsaWater vapor
01:23
SurgeryUniform resource locatorForm (programming)Open setLevel (video gaming)SatelliteSystem administratorAttribute grammar9K33 OsaConnectivity (graph theory)GeometryConservation lawExecution unitMeeting/Interview
02:23
FreewareInformationTerm (mathematics)Reading (process)Gradient descentInformationLevel (video gaming)Multiplication signGoodness of fitOpen setCollaborationismAttribute grammarDiagram
03:13
Electronic program guideOnline helpMultiplication signRectanglePhysical law
03:50
AreaSatelliteDiagram
04:21
SatelliteBuildingMedical imagingLevel (video gaming)Uniform resource locatorOpen setMeeting/Interview
04:54
9K33 OsaOpen setLevel (video gaming)Reading (process)DatenpfadProcess (computing)Complete metric spaceMathematical analysisSet (mathematics)
05:14
9K33 OsaLevel (video gaming)Open setMultiplication signLevel (video gaming)Power (physics)Video gameSet (mathematics)
05:39
Shared memoryEntire functionPower (physics)Set (mathematics)Digital photography
06:08
Mathematical analysisLevel (video gaming)Self-organizationHard disk driveGoodness of fitDigital photographyWeightTablet computerMultiplication signStatisticsSocial classVideo gameOpen setBasis <Mathematik>Lecture/ConferenceMeeting/Interview
07:58
Level (video gaming)Open set9K33 OsaReading (process)DatenpfadAreaOpen setClient (computing)Analytic setFreewareVideo gameStatement (computer science)Bookmark (World Wide Web)Sampling (statistics)State of matterInsertion lossLevel (video gaming)Disk read-and-write head
09:19
Projective planeGodLevel (video gaming)Disk read-and-write headMultiplication signPoint (geometry)Water vaporSampling (statistics)Lecture/Conference
10:08
Service (economics)Water vaporPoint (geometry)Self-organizationGoodness of fitLine (geometry)Level (video gaming)Mathematical analysisSet (mathematics)AreaMultiplication signArithmetic mean
11:14
BuildingPoint (geometry)Decision theoryWater vaporLecture/Conference
11:53
Venn diagramDatabaseGrass (card game)RankingBeat (acoustics)Set (mathematics)Metropolitan area networkWater vaporPoint (geometry)Shared memory
12:44
Source codeSet (mathematics)CircleDecision theoryLevel (video gaming)FreewareClient (computing)DatabaseVideo gameBuildingCybersexAreaSocial classMereologyLecture/Conference
14:52
Context awarenessAreaLevel (video gaming)Open setMeeting/Interview
15:40
Level (video gaming)Standard deviationOpen setMetropolitan area networkMeeting/Interview
16:39
BuildingBoundary value problemLevel (video gaming)Set (mathematics)Different (Kate Ryan album)Meeting/Interview
17:18
Level (video gaming)MereologySlide ruleOpen setLie groupMeeting/InterviewComputer animation
17:41
Level (video gaming)Open set9K33 OsaArea
18:04
Open setLevel (video gaming)Metropolitan area networkMathematical analysisConstructor (object-oriented programming)MetreAreaLecture/Conference
18:49
Water vaporPoint (geometry)Mathematical analysisSet (mathematics)Scripting languageBuildingPolygonAreaFunctional (mathematics)PressureService (economics)Address spaceLine (geometry)Electronic mailing listLecture/Conference
20:02
Water vaporBuildingSet (mathematics)Point (geometry)PolygonMultiplication signLevel (video gaming)PlanningOpen setComplete metric space
Transcript: English(auto-generated)
00:07
Hello everyone And sorry for the title, I hope I didn't hurt anyone Saying that your data is useless. It's not you get good data is useless. It's just data is useless and
00:20
But yeah, welcome some of the slides. I don't know if you Joined our presentation before but some of the slides similar, but the story is kind of different. Yeah, okay Hello So my name is Lydia Okay So I am just going to brief about Geo Gecko, that's me
00:40
Geo Gecko is a geo intelligence company. We work in Uganda, East Africa What do we do geo intelligence maps? Insights, I'm not checking the dictionary but so yes, so this is what we do and as For people who have who work with geo data you geo intelligence or anything similar to that
01:05
You will know that data is at the center of the work that we do and we need data to be able to Do our work. So yeah tonight today we are going to talk about How data is useless in one way or the other by the end, you'll know exactly why we say that
01:26
Hello everyone, my name is desire Yeah, I walk with Geo Gecko working with all kinds of data Name it
01:40
GPS trucks from Rangers in conservation parks satellite imagery admin units Sanitation toilets in campus, so it's so data One data for us at Geo Gecko needs
02:01
Needs to be not only data, but more specifically geo data Data with a location attribute to it. It's not It can't be data can come in many forms, but it's most usable for us when it has some kind of location component Me I guess many of you are familiar with open street map
02:22
Are we yeah, so yeah, it's a collaborative effort From many volunteers to take this imagery digitize it And then people on the ground can add attributes and And that's where the trouble begins as well. Well, it's a note. It's a good effort but
02:45
Again Yes, so there's also challenges with that for example OSM is open street map data is most recognizable through their base map
03:00
But this this this isn't really data this is a This a picture. So if you wanted you could read digitize all this Information, but gladly there's ways to extract that Where you would get this? but It's in they've tried to
03:23
Standardize it in a way with adding tags and and helping guide the people that add data to it to you know help improve the quality of it, but these challenges Most people just do this in their free time. So if I can just draw a few
03:43
Rectangles on the weekend. Hey, why not? So that is a challenge. Maybe laws will expound on but Another issue we have is data gaps most attention is paid to and This is not is paid to areas where there is most activity. This is urban centers. A remote village is not going to have
04:04
That much coverage unless unfortunately disaster strikes and even the same with most satellite companies because they would only collect data for urban centers and along How you call it main
04:22
main roads highways across the country and and so it's potty so Whilst you're using your data If you don't really understand the background to it Hence it's useless. So for example taking this location. This is what it appears as as in OpenStreetMap
04:43
It's a base map This is the data that you have for it, of course, but when you look at the satellite image most of those white spots are buildings and you might think no one lives there, but hey, so The completeness of your data set also matters and that's where I think allows my colleague here can
05:07
Expound upon because we did a geogic or some analysis to Try and see what's out there, especially when it comes to this free available data Yeah, so because like
05:23
We're based in Uganda. So we rely very much on OpenStreetMap data. We can we can rely so much on the government There is a lot of discussion going on with like who owns data If if I collect data, then I own data, then I don't want to share data so so for many government agencies OpenStreetMap can be like
05:45
Yeah, they don't want to go in there because then they certainly don't own the data anymore And they are afraid that they will lose power of the data So if you own the data set then you have the power or at least that's what they think But the problem is that if you don't share data, then it's just useless
06:03
Like there have been a few years ago. There was a company flying the entire country of Uganda Made some really good auto photos and they paid like a shitload of money for that data collection Then all the data ended up in hard drive at the Ministry of Land
06:23
Because they basically didn't know what to use it for but they still owned it then we wanted to like okay should we try to process it so people can actually get some really nice base maps and so on and Based and digitalized from it, but then they said yeah, but then you have to pay for it Because we paid for it. So so if you want to use it, then you also have to pay for it
06:44
okay, so okay we can pay for it and Then okay, but how do you how much should you pay for it? And then there was just like a discussion like yeah, we paid like this million dollars for the data collection, so maybe maybe 50% will be fair and then like it was just like
07:01
Nobody will ever pay like that amount So now the data is just stuck to forever and basically it's a few years old now So now it's useless so they basically just waste a lot of money government money probably funded by the World Bank or the UN or something and And it's basically useless now It was maybe a few person in the Ministry of Land that ended up using it and they probably did something great
07:26
That they know about but we don't know anything about what they used it for and we could have used it for so much Other stuff and we will of course have shared and then then yeah then the discussion starts We work a lot with the
07:40
organizations and people that want to use OpenStreetMap and they have this idea that they look at the data and then they say oh there's Data, it's perfect. We can just use OpenStreetMap We don't even have to ask the government anymore. Then we just make a quick analysis about okay. This is a client that was working in these two municipalities in
08:03
Uganda and they wanted to base all their analytical work on OpenStreetMap because it's free. It's open. It was easy But they didn't know that it was not fully mapped. Someone have played that idea and I heard that OpenStreetMap is complete. It's a complete data set. It's just free. You can just use it for whatever you want
08:25
So this is actually like the green areas is fully mapped But that also means that they could also just be areas where there is nothing So it's not because it's it's actually some sign. It's not because someone has sit had a digitalized something in the green areas
08:42
The yellow areas is like okay someone just has a Wednesday evening and just decided to map one house and then leave the and then left the area and and Then the wet areas are not mapped at all. That was like the example at this I showed before
09:04
And yeah, of course when data is there it's a very easy to explore Python blah blah blah we probably knows the story So yeah, there is data everywhere and it's cool and everybody wants to use it, but we have a problem because There is a lot of people especially in our region that collects a
09:24
Lot of data, but they have this idea in their head that now they own the data They are not so much willing to share it. Of course a human channel sweeper home street map and you in that that are running some initiative and some projects to
09:41
To to to share data, but and collect data and and that's good, but especially in Uganda, I think if you don't Include the government and then get them fully important We'll God just end in a situation where nobody have an idea what they just had to use and
10:04
We have another Says yeah, this is just the example that we have we showed it before in a rua humanitarian home street map have been mapping water points and They are still mapping water points for are they working are they not working?
10:23
That's what they are collecting still and and that's good But in the mean then you have working water lines non working water points and that's good but at the same time minister of water in Uganda, they also have their data set and
10:41
Then suddenly we are in a situation where there's two organizations That basically collect in the same area and they both think that they have the best data set So so there is a mismatch in this data set We don't really we don't really know if this water point is if that's functional or non-functional
11:02
Because they don't seem to talk so much and that's that's a really bad situation when when we are doing this analysis because this is actually people that are using this data to see if they should build a water point or repair water point and
11:20
It's almost a sport in Africa to build water points. It's better to build Than to repair. It's probably easier to get funded but in many situations you could like maybe repair ten water points instead of building one and But if you if you don't have the data to make these kind of decisions then then you can make them
11:44
So it's maybe understandable that so many water points are being built instead of being the repair What else?
12:03
This is just like Yeah, we import and when we analyze and then yeah, I think What you saw before was the Ministries water points and this is the OSM water point So so they are there and maybe some data sharing has happened at some point
12:25
Yeah, so it's not only a problem in Uganda. I think many of us are facing this problem. I'm from Denmark and I
12:45
have also been facing problems in Denmark. We have beautiful, nice, free data sets available and and when you try to import those Then there can also be someone from the street map community that don't want me to
13:02
import one 1,000 building outlines that are that are free, but but it's just like There's there's a problem about who owns the data and I I don't have to answer how to fix it but I see so many in intuitive that where you can share data and you and have a lot of portals, but I
13:27
You take your data from and it's I think in our region is just getting it's just getting really confused so that's also we We tell our clients that the life
13:41
Like we have had a few clients that that we just tell to use them streamer that after as that database and then they can Of course make backup. So if anyone is deleting data Then they can just export again or at least they will have to have their own data set downloaded But it's like to make this circle of data coming in and coming out
14:02
It's just very important, but there is just a lot of issues in the in this In this area, I don't have to answer. I will I will love to hear about what what you are doing in in this regard But I think I'm just afraid that there's so many people that are taking decision based on OpenStreetMap data or government data
14:21
and That that are wrong and that's what that's a challenge that we are facing So yeah data is just useless because there's so much of it there's so many data sources and People don't Especially none. Yes, people don't really know that maybe the data set that they are looking at is probably not
14:47
complete Yeah, I think we will like to some like open up the discussion about your experience with with this area
15:00
Or if you have any question about how we we we work with this I will imagine that it's a problem that you see all over the world. I have seen in Denmark and Uganda and other places
15:26
Yes, hello, yes, thank you, so yeah, I worked in That this region as well in the humanitarian context remote I was never there and we When we did map atons and and mapping around OpenStreetMap we did that under the
15:45
missing maps initiative and I think that is one way but I see at least in a Reform way, but We also had situations where the serious discussion Not talking about any specific country what to put on the map and what actually not to put on the map because certain
16:07
refugee camps are actually you are putting those lives at risk and there were some examples when they would try to attack a Camp because they found it on OpenStreetMap and I think it's also Within the NGO world also NGOs are competing and I'm also wondering what would be wish I would have a solution
16:27
That you won't do it under your own aegis, but together we would agree on a standard way But of course something like infrastructure of course you don't
16:40
there's different kind of data set that you will put into OpenStreetMap and I understand that concern But those that want to see where people live they can also just look at other data sets yeah, map atons is also a good idea, but I think
17:00
For us we kind of make a boundary Like okay, if we may have working in this municipality Then we have to know that all buildings all roads. Everything is mapped in this within this boundary at least Behind you
17:22
Thanks already and superb work there I have a question on one of your slide you showed the level of I say complete and that's not correct like the Coverage of mapping of OpenStreetMap. They say this part is mapped and this part is not mapped Is this something the OpenStreetMap is providing or is it something you computed yourself? I find there is something we did ourselves
17:45
Okay is this kind of work shared somewhere I find this approach super interesting and we have the same kind of problem of people using data Which is probably unbelievable and so Like we
18:00
Went some analysis on like the areas that we could clearly see that there was no reason to check for anything and Then to get the complete picture we also did some of them and the manual work like what you see here is basically a lot of like a Hundred by hundred meters and then so we analyzed every not not all of them manually
18:25
But like we ran some analysis so we can at least for the majority of the data We remove them so we could just say that we know for sure that there's no Construction in this area, but we don't have and it will be interesting to develop
18:42
like methodology so you can actually make this kind of analysis very quick and As I see it there shouldn't be any You could then it must be possible to make them simple Scripts to to want this kind of analysis
19:02
Thank you, but this was just to get the very complete picture of what what what is going on here is data useless
19:22
Everybody thinks that that data set is the best. That's probably also one of the problems But yeah, if you're making and water point pressure analysis in in one of these municipalities then the existing data is useless
19:41
Because you don't get the complete picture You can make that Analysis of water pressure. I actually didn't show that one For those who didn't see this so what this actually show is like Functional and then you have like just simple we will need polygons like service areas
20:01
So it counts all the building outlines within each of the it's just the shortest line between you and the water point and then it creates those polygons and then Then it it counts how many buildings that are within this area, but you can do this in the Rua in Uganda because
20:20
there have been mabotons and We know that the data set is complete but You can also this is based on You can the government data the water points The building outlines are from OpenStreetMap, but there is also water points in OpenStreetMap
20:43
But you you have to use the official data and in Uganda that official data is the Government of Uganda and lucky for us. They said they share that data for some reason and These are easy enough to run like it's it's just Python running behind it
21:01
It's very simple and but you need to know your data So we could do it here, but we wouldn't be able to do it in these two municipalities that you saw before