Project 1 Analyze Week, creating a network dataset for use with Network Analyst Tools to create routes, service areas and closest facility evaluations

Using the information, we created in prepare week, we had several scenarios to evaluate and in order to do so our initial process was to set up a network dataset for use with our Network Analyst Tools.  Our focus this week was creating a map illustrating Evacuation of Patients from Tampa General Hospital on Davis Island, supply routes from the local Armory to three storm shelters, multiple evacuation routes for downtown Tampa and identification of shelter locations. 

      The Network Dataset uses our Street layer containing the additional fields for Seconds, Miles and Flooded which we set up last week.  Within our Network Dataset we established attributes for Seconds, FloodPublic and FloodEmer.  The Seconds usage is cost with seconds as the units and will allow us to determine fast, simple routes for evacuation; The FloodPublic usage is restricted and usage is prohibited which will calculate routes that avoid flooded streets closed to the general public; the FloodEmer usage is restriction with a usage of avoid high which calculates emergency vehicle routes which may need to use flooded streets but whenever possible will avoid them.  It is necessary to set Evaluators for the above attributes in the Network Dataset.  Attribute Seconds, the type was set to Field and value set to Seconds.  Attribute FloodPublic, the type was set to Field and value set to Flooded.  Attribute FloodEmer, the type was set to Field and value set to flooded.

      With the completion of our Network Dataset we began our scenario analysis and mapping.  Each scenario used the Network Analyst toolbar to create a New Route with a logical name matching its scenario.  Tampa General Hospital needed to evacuate patients to Memorial Hospital.  By selecting specific items in the Hospital attribute table, in this case the two hospitals in question, we could load only the selected items into our tool as stops.  The starting hospital (Tampa General) is the first stop and the destination hospital (Memorial) is the second stop.  Because we wanted fast, direct routes we chose to set our Impedance to Seconds and assigned Oneway as the only restriction.  The solve feature was then used to create the optimum route based upon our criteria.  Because of the potential number of patients to evacuate we performed the above steps again but used St. Josephs ad the destination hospital.  This allows for better patient accommodation and also evacuating patients to the best hospital based upon their particular health need.

      Our next task was to determine how best to deliver supplies from the U.S. Army National Guard Armory to local storm shelters.  Because we had three shelters, we need to create three routes beginning at the armory and ending at each shelter.  In this case, we needed to apply both our FloodEmer restriction and the OneWay restriction.  Each route was solved and displayed on the map.

      Scenario three required establishing multiple evacuation routes for the general public living within the Tampa Area.  The closest shelter for this specific area is the Middleton High School Shelter which was used for the destination for each of the calculated routes.  As this is a densely populated area it was important to account for the drive time based upon flooding conditions multiplied by our impedance attribute (known as a Scaled Cost function).  Before preceding we needed to assign scaled cost values to our DEM polygon layer.  The ScaledCost short integer attribute field values were based on grid_code.  Grid_code 0-3=3, 4-6= 2, 7-20= 1 and 9999= null (Null values do not affect the routing process and in this case the value represented water bodies).  By using the values as the multiplicative factor, drive times would be increased based upon the level of flooding.  Now that we have an updated DEM attribute table we were ready to return to our Network Analyst Tool.  Since these residents were all going to the same shelter we chose to create a New Closest Facility Route, in this case Middleton High School Shelter.  The DEM polygon was assigned to the Polygon Barrier.  For the load locations dialog we set the property to Attr_Seconds and the field to ScaledCost.  This was done so that our Seconds were used as the impedance and multiplied by the scaled cost.  The BarrierType default was also set to scaled cost.  Our next task was to load all of our TampaEvac points as Incidents and the shelter as the Facility.  By confirming that the Closest Facility layer properties Impedence was set to seconds, Oneway Restriction was on and Travel from set to Incident to Facility we were then ready to use the solve feature to creating routes from each evacuation point.

      Because other areas had shelters, scenario 4 created service areas and related them to their closest shelter.  This makes it easy for residents to quickly identify their designated shelter based upon the neighborhood in which they reside.  This analysis utilizes the Network Analyst New Service Area feature.  The shelters were added as Facilities.  Within the service area properties, we confirmed our Impedance was set to seconds and Default Breaks value set to 10000; this value represents drive time in seconds and is more than enough time to reach any location within the study area thus preventing exclusion of areas in the analysis if the allowed time were too short.  The Analysis Settings Tab Direction was set to Toward Facility.  Because we did not want our resulting polygons representing the service areas to overlap we set the Polygon Generation tab multiple facilities options to Not Overlapping.  We were then able to solve and produce three defined service areas, one for each shelter.

      With our analysis scenarios performed, we were able to modify our resulting layers to visually differentiate between route types and individual routes within a type category.  For additional clarification, I chose add labels to the symbols for hospitals and shelters.  I also created a DEM_Poly_9999_RH layer to clearly demarcate all water features.  Because the police layer contained both police stations and administrative facilities I chose to produce an additional police admin layer (fdem_policedept_aug07_CPrj_Rh_admin) so that one could visually distinguish administration facilities from actual police stations.