Walkability plugin OS-WALK-EU with Christian Gerten

alrighty so we're live that's great um hi everyone welcome to session two of the um qgis open day thanks all for joining us again for our next um plug-in session um i hope you all saw that there's been an update to the roster at the very end of the day tim is going to have a town hall meeting where we can all discuss the um new release of qgis so i think that'll be great and there's a open jitsi link so if you want to join us in the room that would be great alrighty so on to session two so our session is going to be on the plugin called us walk eu or you and it's a walkability plug-in um at the moment it's experimental of what i can understand it's an experimental plug-in um and it's all about um walkability in neighborhoods and being able to plan infrastructure so i'm going to hand us over or put us into the capable hands of christian he's going to take us through this awesome plugin so over to you yeah thanks amy let me share my screen hope that's the right one i can see it yes perfect it's a presentation great okay then let me have a short look to see my screen yeah perfect yeah hi everyone my name is christian i'm researcher at the ils the institute for regional urban development in dortmund germany and i'm working in the section of geo information and monitoring today i would like to present our workability assessment tool as the os walk you tool it's called and yeah i'm the developer of this plugin but i'm just a spatial planner with some experience in programming and to development and i got a lot of help with from our former informatics student assistants and our technical staff at ils so my presentation is more about showing the methodology behind the tool and i will go also a little more into detail about how to run the tool so i will present a short live demo okay let me tell you about short introduction about the genesis offset tool and why those workability assessments tool are important for us so the ls deals with important topics of sustainable urban development and one of these topics is of course mobility walking has a positive impact on the environment also the social behavior and the health of urban residents and so it's one challenge of the cities in the 21st century to improve the walkability uh to measure this there's a need of walkability assessment tools and currently there are a lot of workability assessments tool um for example the walk score maybe a few of us will know it or the ipad workability but often be seen as not sufficient for our research and therefore we decided to build up our own workability tool always walk you which is based on open cells and open data so that everybody can run the tool by their own and yeah it was a collaboration together with colleagues from university of lisbon into dublin so we worked together on the methodology but the programming part was more part of the ils so the tool always walk you it stands for open source workability tool for european union member states so we can calculate workability for neighborhoods or complete cities and for union member states and yeah currently it's published as experimental plugin so it's an early beta version of of the tool and maybe it's a little bit too early to present it here but we thought it would be a good idea to get some feedback from the experts in the audience and yeah the code is not very complex and just uses native queues algorithm and also cool tool called root service with which we can calculate pedestrian network analysis and yeah currently it's available via gitlab or the cugos repository under the experimental plug-ins yeah now i'd like to go more into detail about the components and the dimensions of walkability which we consider um in total it's five different components which influence the workability in a neighborhood first of all the pedestrian network very basic also the slope population density the green infrastructure and as a supply with amenities or important infrastructures um this is a short or a screenshot of the gui or the user interface of the tool um so just to show you some some some minor stuff later we will see it a little bit better in the live demo on the first tab we have to set the parameters um of the tool for example uh the input grid or input geometry um for which we would like to calculate the walkability which just used input grids especially from the eurostat because there is also information about population which we need for example to calculate the population density um second we need the recreational areas to calculate the green infrastructure uh he will suggest to use um for example osm the osm land use or copernicus copernicus products like um yeah atlas or something else further we need amenities so the location of important infrastructures um i think here it's uh important to use um for example the um data from from osm i think the pys are very very quite good for it and yeah last but not least also the dm so the elevation model here as rester we yeah this is important to calculate the slope in this area on the second tab the user can refine the analysis so we allow the user to set different ratings on the variety of infrastructure for example um the infrastructure is not very easy to understand now i will explain later because um yeah because it's not that easy but also the user can set the weighting of the different dimensions of walkability for example the manatees the pedestrian shed and population density you can set different ratings what is important for the user what is the basic idea behind the tool our idea is to evaluate the workability in the direct residential environment to so to rate the residential walkability um this is defined in the tool as a walking distance or maximum maximum walking distance this could be set by the user we use a default value of 500 meters you can see it on the right side so it's 500 meters over the street network as a pedestrian and yeah mostly all of our indicators are calculated for this area or this walking polygon um as you may receive slope is one of our dimension and walkability but it was not part of the of the gui so we can set waiting for the for the slope this is because we thought we saw that it's not that important to be an own dimension of workability but it also influence the workability so we say you could reduce it's optional to set in the in the plug-in we it could reduce the walking radius based on a simplificating function you can see below so for example if we have in the neighborhood on the right side we have an slope of 30 the maximum walking distance or the default value of 5 meters would decrease to 300 meter so it gets reduced our first indicator or first dimension first component is the pedestrian network schwart called pm that's very important because the directness of routes and the permeability of a street network is important for walking attractiveness so it's basic for walking and uh here we use an indicator called uh pedestrian chat um on the right side you can see now that we have our walking radials and an orange here uh with 500 meters and we have this polygon in relation to a static buffer of 500 meters which for us represents the perfect the perfect breast network and yeah it's just a normal coverage analysis and the value could therefore range from 0 to 100 whereas 100 is a perfect uh walking polygon um for all our indicators so for the pedestrian network for the density the green infrastructure and the amenities we did the reclassification of of the values for example in this case we have in range from one to one hundred percent or from zero to one hundred percent and we classify these values from uh from one to ten as it's more easier to understand lately with where um at which point it is a good value um and that we compare the different dimensions of workability so we did this reclassification for all indicators here i have some examples of how um the build environment could influence those walking distances or the direct residential environment um here's three examples of residential areas in dortmund um the first one um has no barriers um in the direct residential environment or no bigger barriers or for example you see uh is one in the north but it's all many outside the polygon so you can see in the first example that the walking polygon um here in red and the aesthetic buffer is nearly the same so we have a very high pedestrian chat here of i would say around about 90 or something second example represents some semi-permeable um barriers we are really new to the central station in uh in development and but here we can see that there are some minus eq ways um throughout the the the tracks and the train stations so um yeah the walking polygon is quite quite good of the prestige is quite good uh with the value about seventy percent wars um the last example shows some impermeable barriers on the south so you can see a highway which is not reversible for pedestrians on the eastern side you can see a track and on the on the on the eastern side you can see the green a green place a cemetery uh where we do not have in direct access so the excess of to this green green space is on the northern part so we have a very a small walking area which is about 15 or 20 percent pedestrian shed so our next indicator is the population density this indicator was mainly used in order workability assessment tools like in the ipad walkability which was developed and 90s beginners results i think and yeah they use it because they don't have any information about infrastructure or new good data for infrastructure so they say okay high population density indicates a high coverage of of infrastructure supply um but we live in time or in times of google maps and also an increasing and greater ocean quality so actually we don't need this indicator for this propose but also many research shows that the high density allow more social contacts so the people when you live in a very dense neighborhood you have more options to meet people outside for example and this is also an increasing factor for walkability actually here we classified the values in diesels as we do not know at which point at which density the social context or social context could increase so we say okay we classify this into into decile so 10 10 percent of cells with the highest population density gets 10 points up to a lowest 10 percent get just one point um this is a problem because if you just calculate for example the workability for just five or six cells in an area um that could get a problem then we need to say okay this uh indicator should weight it should be rated with zero that does not influence the workability here but i can show you an example later on that the next um dimension is the green infrastructure infrastructure um as this increases the attractiveness of residential environment and also increases and attract active mobility as you all know if you have green green space outside you're more likely to walk and it also increases social interaction so if you have a space a green space outside this is you can meet people for example and in total this creates an healthy environment and also a better living conditions um first indicator we did also some kind of coverage analysis so how much percent of the direct residential environment consists of green and blue infrastructure and so here all the results could also be from 0 to 100 um but here we say okay it's not that important to have for example 100 green and green percentage it wouldn't be good because there's no space for residential and other amenities so we can say okay at a maximum of 22.5 percent and there is no benefit of having more green space so we say then we have 10 points and in 2.5 steps we go down to just one point by less than 2.5 of green infrastructure within the polygon the last indicator and for many assessments tools the most important one is the supply with important infrastructures um yeah here short distances are very important on factor and effective walkability as it's very clear that if you have a supermarket which is 200 meters away it's okay for walking and get other supplies there but if it's not two kilometers away maybe you need a car or other mobility options to get there here we mainly use the walk score methodology but in our case user can select the variety and weightings of the amenities by their own so and also we say that closer facilities are more attractive than uh facilities which are in longer distance um this fact is reflected by our distance decay function we use here it's an adapted cumulative gaussian function you can see so this is a meter and the distance weighting so we say within 400 meters it doesn't it doesn't influence the attractiveness of an infrastructure but then the the curve decreases so for example from if we have a distance of one kilometer or a thousand meter um the distance rating would be zero point five so not that aircraft attractive as for example infrastructure which is 200 meters away um to make it more clear i give a short example here um with two um two infrastructures we have first supermarket here in red um with a rating of five so we rate this infrastructure with five points and as it is very important and everybody needs it two or three times a week you have a distance of the shortest distance to the supermarket 574 meter and another one is the pharmacy in the north which is also important but not that often people need to draw i don't know a few times a month or something so we rated this with uh it was two points and this infrastructure is 759 meters away now we can take our distance decay function and drop on the uh the weightings here or the distance here so for the supermarket we have uh what about 550 70 meters and the distance rated distance would be 0.94 so it's quite very good value for example for the pharmacy it wouldn't be that good it would be 0.77 uh in total now we calculate the total points for for example for supermarket um here we multiply the infrastructure weighting of five by the distance rating it's 0.94 and the final value of 4.7 do the same for the pharmacy which which results in a total value of 1.54 here we can see that on the the sum would be 6.2 with maximum would be 7 so if both infrastructures would be in the distance of 400 meters would be the maximum value in this case it's very easy to understand in which range the value could lie but um if you have more infrastructures and more varieties so if you say okay i look for two supermarkets and uh two doctors also around it would be very confusing so we decided uh to transform the values from range to 0 to 100 so that your user can easily understand in which range um the values could lie and what what um how to interpret the results so in this case um we multiply with correction factor of 14.3 so so the maximum value would be 100 and in our case this in our easy well simplified case would be 89.1 the total result um yeah but in our case we look at uh some more categories or some more infrastructures and total we consider six different um categories of infrastructures which are derived from a literature and also where it's discussed with our colleagues from dublin and lisbon here you can see for example a retail category which includes closes optician books for example entertainment like sierra cinema art center food related stuff like which very important supermarkets marketplace or restaurants civic institutional facilities like college community centers schools but also healthy stuff and office which includes company consulting for example and laptop needs recreation spots like fitness sport center or swimming pools you would also reclassification of values so we say oh we classified into a 10 point classes from um yeah from 100 to to 1. here we need to pay attention that the um that the poi is from oz for example is categorized the right way so we need for r2 we need a new field called type where we said okay fast food is reclassified or is classified as food related stuff night club is for example entertainment and so we provide some kind of translation table for that but the user is also free to do it by their own through say okay for example for infrastructures like bar that's very complicated it's future ledger versus more entertainment so in total the user can decide by their own how to classify those categories okay now that we have all our four indicators calculated we need to sum up all the points um so here we um multiply the points that we have for example for the popular for the pedestrian network we multiply this value by the waiting that we set and degree and that we do it for all or for indicators and then we have a value range which is not which is not very clear as the user can set the ratings by their own so we're also transforming the values from the range from zero to one hundred um let me show them a short example i think it would be much easier to understand um for example if we have an pedestrian chat of 75 we get eight points uh density seven points uh green first drops are five points and amenity nine points um then you can see there's a a screenshot from the gui where we set the weightings of the manatees and waiting for the dimensions for example the manatees in production church full weight population density 0.5 points or weighting and we need infrastructure also 0.8 and now we can fill in all the information that we have so we multiply the rest in set eight by the factor one the weighting one the density is multiplied by 0.5 and so on that we have a total value of 24.5 um it's not very clear now in which range these um this value is and therefore we transform it to make it more understandable so we have 24.5 multiplied by 100 divided by 33. this is the maximum value that we can get if all the facilities would lie in the distance of 400 meters for example and we have the highest value overall so in total we have an uh value and walkability score of 74.2 um this is an example of a complete calculation for a city um the stuttgart in germany um it's calculated on 500 meter bits i think and we also classified these values from very low walkable areas so values up to 30 to a very high walkable areas and 75 and more so you can see here in in green uh the very urban areas um in the city center which are quite quite high walkable also some suburbs are very walkable and in the outskirts more rural areas and sutcut you are very low workability areas so we would say for for these areas um that values up to 45 are more like would say car dependent areas or car dependent enables which are not very walkable and their values above 45 are more likely walkable neighborhoods would say we can also calculate total scores for all cities for example if a shortcut would be around about 60 workability score of 60. for berlin for example it would be we calculated group is 58 and don't want uh not that good 45 workability score yeah and let's go into a short lifetime of the tool um let me share my screen again alrighty live demo i'm ready just as you go over to your um qgis live demo we do have two questions and i think they may be answered in your demo but the first one is from miguel and he asks does the plugin get the amenities automatically and no currently not um we plan some kind of pre-processing um it would be very good if we have that um but at the moment you have to download the pys by your own and classified by your own that is something it depends on do you want to have more freedom for the user or to make it more easy for the user but i think it would be a good option to have that the plugin takes all the pys and classified by their own and then calculated fantastic fantastic i think let's head on over to your live demo and yun will answer your question at the end i think it's a more sort of open-ended question so we'll get to you then okay so then i share my second screen i hope it's visible yeah i can see okay perfect so now we are in qgis um within our master branch we have also some sample data to calculate which are quite here so sample pure eyes a sample grid with also the information about the population in it we have some sample green areas here and also the dem which is not very interesting for the one because it's very flat here um okay then we could start or you also can see the classification of the pure eyes and yeah what is important is that you have the um the root service installed and also set an api key for um for the calculation of pedestrian distances here that's quite important it's also part of our wiki um which hopefully we will publish at the end of the year and we have some problems with our icon also but here you can start the walkability plugin okay here we have input grid which is our sample grid with the field called population here we have some central green here the pys which are classified i can show it before i could ask the question here so here all our f classes are classified for example uh flowers is retail supermarket food related and so on so we need this field type that the tool knows okay how to classify how do i have to classify um this f class okay oh no i started the tool um but it's not a problem i can also um oh no it's some kind of crashing no don't it calculates automatically i restart it but it should be a problem okay um sample guys you can set the maximum walking distance uh you said okay maximum is two two and a half kilometer at the moment and minimum is 250 meters and also the digital digital elevation model to calculate the slope so it's optional for us but um yeah in here so here we can set our now refine our analysis by set the variety and the weighting of the infrastructures so we can say okay how many retail facilities i need did i need three or two or three like as i would say i need three and i can rate the different the different infrastructure so the nearest one it's important for me to get three points uh second and third my lots are important i just get one points and so i can do it for every um every um category of infrastructures currently we have a maximum count of three um we also plan to to make it uh um more more accessible here but currently it's it's a test version of our plug-in um we also can set foot and rings and so on office reset with zero and here below you can set the the weighting of the um of the dimensions of walkability i say amenities are very important for me just get one progressing chat also very good 0.9 and yeah if you say okay population density is not important for me we would like to exclude it from the analysis and set it to zero then we said okay just create a new folder test calculation select calculate now the workability process starts and hopefully it doesn't take so long yeah perfect so now we have our result um we can have a look on the table now so the information population was inside before what we have new is we have a new id which we need for our calculation but then it gets interesting we have our total amenity points so our some kind of walk score here um we have the reclassified point so here would be three so we have 24 points in the humanity calculation we have the population density with the diesels we have the pedestrian shed also classified the green percentage and points and last but not least our result which we could not hear so we can say graduate resort let's take it green to make it look good and here we can see okay we have this classification um let's have a look on the base map to see where this is let's place and don't want our sample data set and we can see here that these two areas are the best one in this calculation step so we say okay these are the most urban one these the other one like like between in some industrial and commercial areas also some tracks here so this um i would say make sense that here the walkability is the best um i brought some more interesting examples so the scapulation took a long so we have here an example of of door mode also um on 200 meter grid um here it takes i think about two hours or three hours to calculate it so it currently it's not very performant to to run this full and with large with a large data set um but here we can also say okay here we have our walkability score result and you see in the center in a city center it's quite good in the rural areas below where we don't have any um or not a good uh supply within first practice is quite bad um and we can can see some suburbs which also have a good um a good supply with infrastructures but also some green areas and pedestrian network what we can also do here is to display for example the uh the indicators here for example the pedestrian chat you can see okay where's the best street pedestrian network you can see for example in the inner center in the city center it's it's quite good as there is the most people walking there and uh or the most ways for pedestrians and in the thousand plants it's um yeah it's not very good and it could be better i would say yeah things that's all from my part i have oh let me get back to my presentation just two more slides i think all righty while you hop over to your presentation um jan just asked how much time you mentioned that it takes some time to obviously run the larger data sets so how much time did it need for you to run the tool for example for stuttgart well i think it would take two hours for for example to try to reach a grid it's not very performant at the moment um but uh hopefully you can see my presentation now um not yet oh let me show ah here ah sorry the right one yeah no um yeah i come back to the major improvements that we have or what we would like to make better that's also the case of the performance um so we are planning to head to an alpha version beginning of next year it's some kind of resources that we need to improve that tool but yeah some minor changes like simplify the code we have some redundant parts in it that we need to improve and proof also the gui look it's so important it doesn't look very good at some off um you need some time to make it look better and yeah current challenges i think that's the most important one um we need to increase the performance as currently we calculate we can't calculate the closest uh facility or manatee at the moment so we say for example in our tool you can say okay i need three the three nearest facility but the ors has not this function so currently we need to calculate the distance metrics so from the point to all amenities in this area and that's quite of resource ads it took a lot of resources to calculate it and the next problem is that we have or the os does provide just the limited uh requests per day so limited token i would say so normally the user wouldn't be able to calculate a whole city because it would be out of range out of the free requests per day we have installed an own version of the ors at our server so this works good but i think nobody can not everybody can do this so our plan is um to simplify our distance rating function so it doesn't that we do not need uh to calculate the direct distance that we make it more like looking for steps so simplify the function and then simplify the code in just using the native cubist algorithm function count pointer polygon it would be [Music] very easy to calculate it just to calculate some more uh treble time rings or walking rings and we would much i think it would make it much easier or much more better better performance for it and yeah some more ideas are for us to define profile settings for example for elderly people that we say okay every people do not walk such long distances they need other infrastructures that we could define profile settings offer also for younger people and another idea is to integrate for example public transport data from gtfs files like frequencies or transport type to identify car dependent areas or really car dependent areas and as i showed in stuttgart we define the lowest categories of walkability as car dependent areas but total we don't know because we do not have information about other mobility options like public transport and that make it more easy to to see the structure or the mobility structure of a city um but this would be more kind of another branch of the tool so maybe maybe a second tool would derive from that okay some kind of last conclusion uh we are the early stage of development i would say um and maybe it's not that easy to follow the methodology and therefore we need or we prepare wiki with detailed information about the calculations and so on and yeah we also plan to publish the scientific paper or we also wrote the paper and i get a review for him for it and hopefully we published and of the year beginning of next year which is also who could be some good reference to understand the methodology behind the tool and yeah as i say an updated version of the tool at the beginning of next year hopefully and yeah if you have any ideas questions critique or ideas for collaboration please ask or contact me and also by mail and yeah um also you can see some our web trs which is our city region monitoring but it's just in german so maybe not interesting for all for us but it provides some nice maps and we also include our workability indicator in it so if you uh want to have a look on it just uh go to the website and yeah if you have any questions ask and yeah thanks for having me brilliant stuff thank you so much and thank you for those um contact details um i think this is such a fantastic plug-in and it's got so much um possibilities you know um and is very very usable for you know infrastructure planning and infrastructure use i think one of the main questions i had at the sort of very beginning um when you mentioned that it's mainly for european countries um is is there a plan to expand into other countries or is the application able to um apply to other regions currently it's possible to calculate for every city you just need the data sets so for example if you have a digital elevation model for other cities it's no problem so currently our research focus lies on european cities and so that we say okay we just focus on european cities but i think also the tool is usable for every um for every city it's as as it's good data from for example osm and and so on so i think the only issue is currently is the uh the data quality so if you have data that's good for an area then it's possible to calculate of course absolutely absolutely and then that last sort of more open-ended question and i think in a way you did answer it but i thought i might come back to it just in case there's some more interesting news um and that is from jan and he asks um what are the plans for development of the plug-in so in other than of course yeah yeah so as i said i think the most urgent one is to increase the performance of the tool to make it more uh better performing and um yeah hopefully we have refined or i find the time to develop it further end of this this year um to make it more performant and um yeah to make it more good looking currently we have some some issues that for example the ors plug-in changes so we also have to look that our tool works when this plug-in changes so um yeah we do not have so much resources to to develop it further so it's many of my shoulders to to to have a look on it and yeah so we do our best and hope we can can finish an alpha version beginning of next year to make it more and make it uh run better absolutely so it really is sort of the the bleeding slash cutting edge of the development of the application and i think it's going to be great to watch over time where it goes and how it evolves i always find that sort of thing really exciting um so if anyone else has any other questions please post them now or else i'll start wrapping up um this really really cool session on um walkability i think my last question to you chris is where did this kind of come from so what was your fir like who had the idea this is what we're gonna need i know um particularly in europe there's quite a movement towards you know more walkable areas um car free zones and cities was that generally like the spark that ignited making an application like this or where did the like base idea come from yeah i think it's it's a really great debate in in europe or in germany also that um for example um i just like the 50 minute city and so on that can everything within 50 minutes i think it's it's great great idea behind it and many also now are in our house in the ils there are many researchers who deals with walkability and they ask oh can you calculate that for me for example and there was some kind of point we said okay we it's good so many people need this as a staff or need data sets or need such analysis that we said okay we could develop our own tool uh we had some also some great collaboration with with two dublin and lisbon that we can say okay with so so not so much knowledge about the methodology behind uh calculating or assessing walkability that we could need or we could use them to develop our own tool so was just write down all the code stuff and hopefully and hope that it works and yeah now we have the point that hopefully it could get better and yeah brilliant stuff all right so i'm then going to wrap us up there and i just it leaves me to thank you um christian for coming and giving us an amazing look into the tool that is being developed and is does currently have a working version and i hope that you know early next year or in the next year when the tool is up and running and all those you know little kinks have been ironed out you can come and give us a session just showing the updates and how awesome the tool is working so thank you very much yeah thank you i hope looking forward it and hopefully it will happen fantastic all right cheers everyone um please log in in the next hour top of the next hour for the next session and i will see you all there thank you

Channel: QGIS

Views: 332

Keywords:

Id: sOJoNhh4E90

Length: 42min 0sec (2520 seconds)

Published: Fri Oct 29 2021