The spider aids urban designers in analyzing and understanding varied user behavior as well as generating a basic framework based on quantifiable data. Spider explores the intersection between artificial intelligence and human intelligence to find a balanced workflow between them. The tools facilitate three stages of the design process Analysis, Generation, and Visualization. The analysis part deals with the correlation of contextual data based on human behavior which includes view, user movement, Radiation, and wind movements. It also deals with bringing different sources of available data to a similar format for a seamless workflow. The generation part deals with creating rules based on the available data and producing a basic framework for designers to work on.
(Please navigate to the respective pages by clicking on the images below for the further details about the models)
Plot generator
The plot generator is a tool that creates volumes by dividing the given urban site. The model splits the volume in a branching fashion based on the depth given. This can be integrated with the other generator to create various organic plots. The outcomes are then evaluated with optimizers to be altered by the users intent.
Grid generator
The grid generator is a tool that creates volumes by dividing the given urban site in the form of grids. The model creates a grid based on the width and length which is used to split the given site. The outcomes are then evaluated with optimizers to be altered by the user's intent.
Analysis grid
This tool is to create point grids for spatial analysis. The model generates a point grid with the given size and collects the points within the boundary. The output is used for multiple analyses like thermal, energy and carbon footprint.
Isovist
The tool uses a ray tracing algorithm to find the isovist diagram for a given position in the context. The isovist can be used to describe space from the point of view of a person within an environment.
Shortest distance
The component finds the shortest path between two destinations in the given road network. The tool uses the Dijkstra algorithm by analyzing the spatial graph from the network. The given network is first trimmed of loose ends. Then the distances are used as weights. At each step, the model evaluates the total sum of weights and chooses the lowest sum for the shortest path.
Concave and convex hull
The Concave and convex hull algorithms use the Jarvis march algorithm to find the bounding region. As a primary step, the tool finds the lowest point in the cloud. The next point is found by having the largest angle in an anticlockwise direction. To find the concave hull the whole process is applied for a selected number of points near the initial point.
kd Clustering
Here we adapted the kd clustering algorithm to subdivide the given urban neighborhood into plots of given volume. The algorithm was efficient to be used on varied scales. The algorithm starts by initiating the clusters with random points. At each level of progression the model allots the points to the nearby clusters. Once the cluster reach the predefined weight it is are taken out.
Offset Curve
The tool uses minkowski sum to create offsets for the given polyline. The input is converted to a array of points and a circle is created with a given radius. The Minkowski sum is found by the convolution followed by decomposition of the inputs.
Polygon Triangulation
The polygon triangulation is mostly used to subdivide the complex shapes into basic triangular units. The ear clipping algorithm was incorporated for this function. The model computes all the interior angles and the vertex with the smallest interior angle is used as the ear tip. The triangles are formed when the ear tip doesn't cover any reflex vertices inside them. The process is repeated until split the whole polyline.
Split with line
The model uses an edge tracing method to split the given polyline. The polyline is converted into an array of points arranged in the anticlockwise direction. The algorithm then traces each edge for intersection and splits the edges. Resultants are closed to form separate polylines.
Get image location
This tool reads the location data embedded in the images we capture and uses Mercator projection to plot them in the digital space. Primarily the tool is used for accurate site documentations. The tool can be used along with the OSM reader for precise mapping.
Read OSM File
The tool reads OSM files and using Mercator projection to plot them in the digital space. Each feature in the metadata can be accessed separately higher degree of freedom. The OSM files store all the features in the form of tags, attached to its basic data structures (its nodes, ways, and relations). Each tag describes a geographic attribute of the feature being shown by that specific node, way or relation.
Minkowski sum and difference
The Minkowski Sum and the Minkowski difference models are prominently used for robotic path detection and efficient packing. Both the inputs are converted into array of points and the model uses an convolution algorithm for both sum and difference.
Polygon Boolean operations
The Edge tracing algorithm was adapted to make Boolean operations with the input shapes.