Hi Adam, I had mentioned that I downloaded data of Indiana and but the code for converting it to a routable network has been running for almost 90 hours now and am not sure how long it will take.
So I have now extracted osm data of Indianapolis city only and have started the code for converting it to a routable network. Once this code completes execution, I will upload it to the server and focus the map to Indianapolis. Thanks, Nadeem On Tue, Sep 17, 2013 at 4:00 AM, Adam Estrada <[email protected]>wrote: > Awesome job, Nadeem! I'm pretty sure that you could have used Osmosis or > the like to extract the area you are interested in but we can save that for > another project :) You're almost there so keep up the good work. One more > thing though...please change the map extent from Taiwan to Indianapolis to > match the OSM you have downloaded. > > Adam > > > On Mon, Sep 16, 2013 at 4:39 PM, Nadeem Anjum <[email protected] > >wrote: > > > Hi All, > > > > The update regarding the project is as mentioned below: > > > > *Saving Data as GeoJSON* > > > > - I have added a function that converts the map data into the > standard * > > GeoJson* format and saves it on the server > > - I have added a function that converts the saved geojson file and > loads > > it onto the map > > - A part of the non-spatial data like schedule etc is being saved in > > json format as a separate file > > > > *Behavior of Distance Decay on Map with additional conditions* > > > > - A part of my GSoC proposal included viewing the results of distance > > decay when simulated on a map with conditions like presence of police, > > presence of guardian at crime spot, inherent risk, profit value of > spot > > etc. I have implemented this as follows: > > > > > > 1. The agent starts from "Home" and has an initial equity (set to 100 > > for now) > > 2. A number of crime spots are marked on the map, with parameters > > similar to those used for simulating routine activity > > 3. Agent takes "Random" form of movement > > 4. Agent loses one point from the equity from every km moved > > 5. Agent returns home when equity is running out > > 6. Agent commits crime, based on similar algorithm as used for routine > > activity. This includes inherent risk, profit value, number of times > > visited, presence of police, absence of guardian > > 7. If the break-in is successful, the profit value is transferred to > > agent's equity > > 8. If the agent is caught, the agent loses 25% of its equity > > 9. Probability of an agent being caught is proportional to the total > > risk > > 10. After the break-in, the agent returns home and restarts the > > simulation with new equity > > 11. This repeats until the equity is reduced to zero. > > > > To run a simulation, > > > > 1. Go to: http://gw76.iu.xsede.org/criminfo/ > > 2. Click on "Mark Home" and mark it. There is no need to mark other > > locations like office etc for this > > 3. Mark any number of crime spots you want > > 4. Click on Simulate Distance Decay > > > > A sample data has been saved with home and a large number of crime spots > > marked. You can click on "Load" and enter the name "dd" to load it. Then > > click on "Simulate Distance Decay" to simulate it. > > > > A sample result of the simulation run 4 times has also been saved as > > "dd-result". You can load it to view the results directly > > > > *Changes to UI* > > > > - To avoid cluttering, only the number of successful break-ins are > being > > shown in the labels of crime spots now, Other parameters can be seen > as > > popups on click on the crime spot. So crime spot label is CS-n [f] , > > where > > n is id and f is no. of successful break-ins. Pop is [x,y,z,t], where > x > > is > > reward, y is total risk, z is probability of police being in vicinity > > and t > > is probability of guardian being away. > > > > *Loading Data of Indianapolis* > > > > - I did not find osm data of just Indianapolis, but I found that of > the > > region Indiana. ( > > http://download.geofabrik.de/north-america/us/indiana.html) I have > > downloaded the whole data of Indiana. The code to convert it into a > > routable network has been running for over 3 days. I hope it completes > > soon. Once it does, I will upload in into the server and we can run > > simulations in Indiana, without making any changes to the existing > code. > > > > *Patch* > > > > - I have submitted a patch to the latest code > > > > > > Thanks, > > Nadeem > > >
