cityscope / csl_guadalajara Goto Github PK

Add a digital Elevation Model using this model as an example in the GAMA model library
Data/Data Importation/models/Grid DEM.gaml

Check the Levy function whether it makes sense or not to use it (vs getting the nearest interesting cells). Right now some target seems too far

Recap from @LeticiaIG

1.- How does individuals’ perception of safety change their behaviour?
1.1 Starting to approach the issue thinking about what they are doing now.(They are going form home to work)
1.2 Do they avoid certain places or there are places they cannot avoid?
WHAT IF scenarios: as an agente, I check my perception What if I change something Would I behave differently?

2.- Visualization

    A. FORMAL SURVEILLANCE
            Play a bit more with dynamic. Now it is 0-1 > (change) Distance from this car
            It is cool to have difference behaviour, but it should be more obvious.

    B. NATURAL SURVEILLANCE
            Same thing that A

    C. MAINTENANCE
            OK

    D. PHYSICAL ISOLATION
            More differences in the colors
            Better way to show the ranges

3.- Frame the validation process

@LeticiaIG has the idea to add some facade

You can take some inspiration from the GAMA model library in Visualization and User Interaction/Visualization/3D Visualization/models/Procedural City.gaml

to have something like this

aspect textured {
  draw shape texture:[roof_texture.path,texture.path] depth:depth;
}

@gamalielpalomo I am unfortunately not able to run it with the last version of GAMA.

Would it be possible for you to have a look it the version in the master is effectively the running one?

I see for instance this part of the code not being able to compile. Would be great to have a working version if possible (or maybe it living in another repo in this case would it be possible to link to it?)

file roads_shp <- file(dcu_roads_filename);
file ppdu_blocks_shp <- file(ppdu_s2_blocks_filename);
file transport_shp <- file(dcu_transport);

@gamalielpalomo

What is the point of having a species suburb if it is only use for display ?

Discussing with the team we agreed that @LeticiaIG and @gamalielpalomo will integrate some of the safety and security indicators in the ABM Model

A first version of the documentation can be found here

https://github.com/CityScope/CSL_Guadalajara/wiki

Using the parameter sliders let's have a dynamic mu sliders so the user can play with it

• Possibility to load any shapefile from the interface
• Compute another heatmap that will show the distance to the hospital (so there will be 2 displays)
• Produce the transport network for Miramar and Tijuana

We need to validate the simulations with data collected from polls and Leti's app.
Right now we are thinking in two validation techniques: the parameter sensibility and the comparison with real data.

In order to compare real data with simulations results:

1. Adjust the model to produce the same kind of outputs as the available real data.
2. Select different areas/places of interest in Lomas: Camino a la Mesa, El Bosque, El Arroyo, El Cristo.
3. Compare the perception of insecurity of simulated agents passing through each of these places, to real opinions.
4. Produce an accuracy value for the results.

Using the overlay facet we need to have a legend in the display to explain what we are seing.

• Add legend to Identify the uses for people
• CROP the model now
• Clarify the narration
• global perceived security
• Output > (ASPECTS) > go inside these aspects
• INPUT > OUTPUT

To save a bit of computation and also being able to see better the studied area, the model need s to be recropped to correspond to the DEM envelopes

Right now there is a global perception value computed as a weighted mean of the 9 indicators.

We should explore different visualization using 9 displays for instance in a different experiment

We need a place where we have a very simplistic description of the model. The wiki is a good starting point

Aspect

• Create one aspect to trigger each parameter by pressing 1,2,3,4,5 and parameter on the left
• Create one aspect with every parameter using cylinder

Social Graph

• Use one color per family
• For each agent add a user_command to be able to click on the agent and show its family, friends,

Building

• Add the key building (church, school) in the .shp file by adding an attribute type and defining the color in GAMA

• Perception
  I would keep the color for a global view of the map
  And maybe let's explore something else for individual

• Play with sunlight by changing the background of the display (just keep it as a boolean parameter)

Every n steps, each agent recomputes its perception value. This makes the simulation to freeze at that point when we use a considerable number of agents. How can I update indicators values for each agent and not to have these pauses? I tried with paralelization, but it didnt make difference.

reflex update_indicators_values when:every(10#minute)

I think it would make sense two create 2 different species in order to be able to differentiate offender and victim that in long term might have very different behavior

It can be nice to have an aggregated heatmap to see where the crime occurs more. Maybe a third display we each time there is a crime you increment a value for a cell

Need to find a way to display the global perceived security, by using some charts but also at the individual level

Ones could imagine the user can click on an agent and displays its own security perception.

This can be achived by using a dedicated aspect