Latest Contributions by JaySandhu

You will first need to convert the polygons into "points" to make them locations of a network. You can do this either by replacing them with a centroid location OR if the park/green space is big, find the intersections with all the roads around them. Once you have these points, you can use the OD Cost Matrix solver (or the Closest Facility) and load your green spaces as Destinations and your bus stops as Origins and set a cutoff of 25 minutes and solve. The resulting Lines sub class will have the distances to every green space "destination" and you can Summerize this table to get number of locations per bus stop (origin). If you need the shape of the actual paths taken then you can do the same with Closest Facility, loading the bus stops as incidents and green spaces as facilities and setting a cutoff as well as max number to find equal the the total number of facilities you loaded. Jay Sandhu ... View more

You can create a network dataset in a wide variety of ways. It can support end point connectivity (default) or any vertex connectivity. For end point, it assumes that street feature with a common end points will meet and you can transfer from one to another. You can override that with additional information such as elevation fields to indicate that there is a bridge or tunnel there so that proper connectivity is created. Without this additional information you cannot create correct connectivity. TIGER data's goal was to support the US census activity not provide the best routing data available. You will need to either enhance it with additional information or find another source such as open source map. However if you are planing to use TIGER, why not use the free streetmap data that ships with ArcGIS. It is on the Data and Maps DVD and covers all of North America. It is ready to route. It is not the most current data (i.e., it is free) but will be better than TIGER. Jay Sandhu ... View more

First I would like to say that you can control where the points load by using a query. That is, at load locations time you can specify that the point should not locate on highways or ramps. To do this you use the Build Query option where the query could exclude roads with say CFCC code A63 etc (based on your data). You can read more about it in the section titled "Snapping environment with Build Query" here: http://resources.arcgis.com/en/help/main/10.1/index.html#//00470000003n000000 Another way to generate candidate facility locations (that I like to use) is a combination of two GP tools. These are the Feature To Point and Create Random Points. So you can start by selecting a set of roads that meet your criteria for locating a facility and then use the Feature To Point on this selection. It will create one point per street. If this is enough you can load them as candidate facilities. If they are too many or too dense, use the Creaet Random Points to further subset or randomly select from these points a set of points that are say every 1000 foot apart (and you can specify how many you want). Just remember, it is better to logically narrow down the candidate set of facilities so that you can get a good feasible location. So if many of the roads/locations are not available for building a fire station then exclude them before you run the Feature to Point tool Regards, Jay Sandhu ... View more

Most likely there is a high cost street in that area. Bring up the properties of the service area layer and turn on the Lines and have it split lines at breaks, re-solve. Now examine the lines in the area of interest to see what is the cost assigned to the edges (from/to cummulative costs). That should point to the underlying network costs causing a distortion in the polygon. Jay Sandhu ... View more