GIS5100 - Module 1 - Crime Analysis

                 In this first module 1, we learn how to hot spot analysis using grid overlay, kernel density, and Local Moran's I. This first lab is definitely a step up from the previous classes, There were a lot of trials and errors for me to self-experiment since the directions were more vague and I had to relearn a lot of the tools we used. Belows are a quick summary of the steps I used which I hope prove to be useful to others as I personally struggled a lot with this first module. 

               I created a blank map called Burglaries and added data from the WashingtonDC dataset. I created a new table using “Select by Attribute” for ‘Burglary’ from the “OFFENSE” field and named it “Burglary”. I then used a spatial join to count the number of crimes per census tract. Next, I joined the generated “Burlaries_Tracts” w/ “dctract_data” using “TRACT” and “Tract Number”. Then I added a field to the “Burglaries_Tracts” named “BurlgaryRate”. Next, I calculated the burglary rate using the provided python expression. I added a graduated colors symbology which excluded the burglary rate of 125 or 1500 from the provided SQL expression. I then created a layout and added a legend, then exported the layout as a JPEG.

    I created a new map called “Assault” and added the “CrimpeRep2018_DC” and “Boundary_DC” to the map. Next, I select by attributes where offense is equal to assult w/ dangerous weapon and created a new layer with the selection. I then open the Environment setting and set the extent to the Boundary_DC and mask to Boundary_DC. Next, I ran the Kernel Density tool with the provided inputs. I applied a Classify symbology to the generated assaults_kd. I excluded 0 from the layer’s data. Then, I set the upper values to the required mean plus means * X, and labeled it accordingly. I then created a AssaultRate layout and added a legend and exported the layout as a JPEG. 

    For the grid-based hot spot, I spatial joined the Chicago_Grid and the 2017 homicide which made a new homicide_grid layer. The homicide_grid layer now have a Join_count field which I use select by attribute to select any rows with higher than zero. I exported all the rows previously selected to a new layer call “homicide_count”. I then open the “homicide_count” and sorted the “Joint_count” from largest to smallest and then I selected the top 62 rows and then exported the 62 rows as “hom_top20per”. Next, I added a field called dissolve which use a short interger. I added the number “1” to the entire column. I then dissolved the layer using the provided settings to make “hom_top20per_diss” layer. 

            


           

    For the kernel density hot spot mapping, I used the provided inputs in the instruction and modified the symbology so that there are only two break values: 3 * mean and the maximum value. I also excluded zero from the symbology as well.  I then reclassify the raster and used the provided inputs and exported “hom_kd_rc”. Next, I converted the rasted to a polygon and named the feature class “hom_kd_fc”. Next I selected by attribute for the gridcode value of ‘2’. I exported the selection as “hom_kd”.

            



    For the local Moran’s I Hotspot Mapping, I spatial joined the census tract and the 2017 homicides and named the output feature class hom_morans_sj. I then added a new “Crime_Rate” field with a float data type. I use a python expression to do calculate the homicides per 1000 housing unit using the join_count and total_households. Next I ran the Cluster and Outlier Analysis spatial statistics tool and named the output hom_morans_run. After, I used SQL query to select the high-high cluster and output the selection to a new hom_morans_hh layer and dissolved it. 

            

            For the table, this is the result I got for others to compare.
          


Comments

Post a Comment

Popular posts from this blog

GIS5007 - Orientation - About me

                     Hello everyone, my name is Quang Nguyen. I graduated from Florida State University in 2023  with a B.S. in Geology. I currently work full-time at the Florida Department of Environmental Protection in the wastewater compliance assurance program. I was able to use the FDEP tuition assistance program to pursue a certificate in ArcGIS at UWF. I choose this program because of my background in Geology. I believe GIS will have a important role in the future of the world and I want to become a part of it as I transition from my current job to a more GIS-focused role in the future. I hope to gain a lot of skills after finishing the certificate to do more environmental work that uses my Geology degree to do good for the world. My story map below shows my journey since I was a little kid, and how rock climbing allowed me to travel across many states of the USA. I hope one day I can combine my passion for rock cli...
Read more

Final Project - GIS5050 - Bobwhite-Manatee Transmission Line Project

Image
  This final project was an accumulation of all the skills we learned in this class. I had to use a lot of skills from previous modules that was easily forgotten. I found that creating the layout was one of the most challenging tasks consider what I have learned about neat lines and things such as placement of legends, or scale bar divisions. The above map is a map with all the layers I generated for this project! It looks very cluttered but honestly the parcels part of this project was a very hard task to work on. I also decided to use arcgis story map instead of powerpoint for the presentation which added a lot of extra time to learn. I finds that story map UI is not very user friendly to uses when creating textbox compared to powerpoint. I want to go over the 3 main maps I made for this final project in this blog post and my overall process/experience with them. The environmental land map was a good introduction on how to use the clip tool to isolate the conservation lands/NWI l...
Read more

GIS5007 - Module 4 - Data Classification

Image
           This module focus on learning the 4 common data classification methods: Quantile, Equal Interval, Standard Deviation, and Natural Breaks. By using one dataset but 4 different classification method, the dataset can be visualized differently. Below is how I interpreted the 4 classification methods.           Equal Interval divides a range of data values into equal-sized intervals (based on how many classes are given). It is excellent for uniform distributed data where values are spread out evenly. Extreme values are placed in the lowest or highest class. This can hide some values if some are not evenly distributed.            Quantile classifications divide a dataset into classes with the same number of data points, regardless of actual values in those data points. This ensures each class has the same number of data points, which can be beneficial for data that is evenly distributed. ...
Read more