Grasshopper, Isovist, Urban Analysis

In this tutorial, the idea is to explore the potential of Isovist for urban analysis. Isovist is the surface or volume of space visible from a specific point. This concept has been proposed by Clifford Tandy in 1967 and then redefined by Michael Benedikt. Isovists are very useful to quantify the perception of urban spaces such as opening, closeness, and also useful to define urban envelopes. Isovists help to with non-intuitive solutions for complex problems.

In this example, we will see how to set up an environment, then use a simple Isovist System, then a more complex one with Ladybug. Then we will see a non-intuitive question: where a very high tower can be seen in a dense urban context.

Let’s dive.

Importing OSM data.

Depending on locations, OSM data can be exhaustive with roads, 2D buildings, 3D buildings and many more information. OSM are data that can be imported from Open street map. www.openstreetmap.org

In Openstreet Map we choose our area of study.

In Export we choose the area of study with manual export.

We get an map.osm file.

We rename it with the name of the location or area we are.

Grasshopper import

Grasshopper can import OSM files, but a new plugin is needed. This will be ELK that can be downloaded here: https://www.food4rhino.com/en/app/elk

Choose Elk 2.2.2

To install, just drag and drop all elements on the grasshopper window.

We now have the Elk Components

Set as followed: we prepare a File Path to import the OSM file that feed the Location Component.

We connect OSM data and connect OSM Point Data and OSMFile

Nothing appears.

We right click on the component Icon and click on Create 3D Buildings

In our case, we are very poor in 3D buildings. Too bad.

In this situation we’ve got the tower that is what we will study. We will have to filter the data to find and isolate the tower

It is possible to number all shapes using this structure.

Flatten the building output. Add a list Item and identify the tower. Here it’s number 166

We will Cull the tower so that we will have on one side, the tower and on the other side all other buildings.

Extruding the buildings.

We will consider that all other heights are similar.

Ok the context is ready, we can start to study our Isovists.

2D Isovists for a single point

For more clarity, we organise our definition as this:

We this first the simple Isovist function which is Intersect slot, Mathematical collection

We create a point which is going to be the location from where we observe, Count is the number of Rays, radius, the max distance we observe.

To visualise the results, we will create a geometry to connect all resulting points

To better see the point, it is associated with a sphere (green in the image). A polyline connects all points. The polyline has to be closed. Then a surface can be done out of it.

If we move our point we will see the different Isovists

It can also be used for cut sections

Just add a XZ plan or YZ plan or whatever vertical plane

We can exclude all points under the ground by a simple sorting process; any point that has a Z coordinate superior to 0 is kept.

Better.

3D isovists

Now we want to do the same in 3D. We will use the Ladybug components. We will use the LB visibility percent.

We will investigate how much the tower is seen.

The previous Isovist system will not be used anymore. We just keep the point. Then we connect as shown…context is context and the tower wil be our geometry.

We click on the Boolean Toggle to start the simulation

This is how the Tower is seen from this location

If we go in a narrow street…the tower is no more visible despite the fact that it is super high.

Exploring along a path

Now we could try to see what is happening when we walk in the street, how often we see this huge tower.

We create a line that we will divide by distance, saying we test every 5 meters, or more or less….

Let’s try

We can try to see for every step how the façade is seen, we can draw the lines of sight.

This is the visibility for point n°2

This is n°3

N°10 as 6,7,8,9 don’t see nothing

To prepare that, we just sort the points with 0 visibility from the others…

Mapping the visibility

Now, we will try to have a global view of the visibility from the path itself. It means that we want to see with a colour code when we do see the tower and when it’s masked. To do that, we will affect to the circles a specific colour, for instance, red, no visibility, and yellow partial, blue full visibility. Blue will never happened as it is not possible to see all sides of the tower at once.

Ok, the best score would be 100. The chosen grid size gives a certain amount of faces to be investigated.

Now we have 28 points on our path and 472 spots to test on the building.

We have to test every point on the path, get the result, and transform it into a colour. For that we need a domain. We associate a gradient of colours to a gradient of values. 1 number = 1 colour.

We graph the flux of points, so that every point on the path will be tested and then the next and so on.

We extract the number of points tested on the tower

But as we graphed, we must take the first iteration only, not the whole repeated process…

Then we will create a domain. Upper limit, the best score possible, lower, the worst and our results will be in between.

The best theoretical result is number of points on the tower multiply by 100. This is the domain end (best score). The domain starts with 0.

Ok now we’ve got our max and min numbers. To make things simple, we can remap this domain from 0 to 100. Like a percentage.

Now we just have to input the calculated values.

In the results, we have the score for each tested point. So we add everything (Mass addition). And we connect that to value. Attention, Mass Addition result has to be flatten.

Let’s check:

Ok, now we affect a colour for each value according to a gradient.

We set a custom preview, input geometry will be the circles, and colours from gradient. For each point from the path, we check the value and the correspondent colour.

Now we see the result

To have more details, we can test more points on the path, every 5 meters

We can also decrease the Upper limit to have more nuances.

Investigating a Surface

We can replace the path by a surface to investigate how much the tower is visible.

This is how is define our surface, it’s a simple setup for the test.

We connect the points to View_points and we add an offset to fit with the eye level

Result, the grid is large, (7 meters), we can try to get more details, but it will cost computer power

Last step, open visibilities.

The perception of space can also derive from how far is it possible to see. Is density compatible with wider vision? We will see how to quantify this aspect using the first Isovist component we saw above.

We will create an Isovist for each point on this grid

We reuse the same definition

Connect every point of the grid to the Isovist definition

Now we could calculate the distance from each point to the first obstacle. The less it is, more confine is our space, and the more it is, more opened is our space.

This is what we got.

To make it more comprehensible, we will decrease the number of rays, and hide the surfaces and then create a segment for all rays.

Now we will calculate every segment’s length and add them for each point.

We’ve here 170 points so 170 values

Ok, same as above we will create a colour from a gradient for each value.

We can reuse what has been done before.

So from the results, we look for the Min and Max value to create our Domain. We flatten the results and then look for the bounds and it gives our domain.

We remap our number for a more comfortable use, and input the Values.

Then we connect to a gradient

Then those colours are distributed to all faces from our surface of study.

It’s done, this is our openness visibility study

Laisser un commentaire

Votre adresse e-mail ne sera pas publiée. Les champs obligatoires sont indiqués avec *

Ce site utilise Akismet pour réduire les indésirables. En savoir plus sur comment les données de vos commentaires sont utilisées.