suggestions for Isaac and Chris
The goal is to create polygons (not lines) for areas around major street, so that you can capture (or "join") the crash points that occur inside those polygons. The result will be a spreadsheet (CSV) of crashes joined with polygons that displays a new column named corridor, which includes the name of each major street that you added as a polygon. Using a pivot table, you can view how many crashes occurred in each major street (if you added them as polygons above).
This repo contains files you sent me, plus new ones I created following the suggested steps below. You can view each by clicking on the file name inside GitHub in your browser, or go to Code > Download as Zip to work with them using other tools.
- You sent me two files (10streets.geojson and hartford.geojson) which I placed in this repo. When I placed 10streets inside https://placemark.io (click use for free), I see it is only lines that made to appear to be very thick. But these are not polygons.
- Fortunately, you can use the https://placemark.io tool to buffer objects, which converts lines into polygons. (This is less precise than hand-drawing each polygon, but it's definitely faster.) I buffered by clicking on a line (Farmington Ave), then right-click to Operations > Buffer. Next, I need to set the width of the buffer around the line to convert it into a polygon, and I arbitrarily chose 20 meters (you may choose differently). Click the Buffer button.
- Now see how your line is now a polygon, and is labeled correctly in the table of contents, too.
- Now that we have a polygon to represent the street, the next step is to properly name for when we eventually join them to crash points. In the Placemark.io tool, see the right side Feature, where I created a new property named corridor, and inserted the value Farmington Ave.
- Next I used the buffer tool to convert a second line to a polygon. On the right side, notice how I clicked from the Feature to the Table tab, which allows me to easily enter the value Main St and easily see all corridor values. Also, I deleted the unnecessary properties (line thickness, red color) to avoid confusion.
-
In the Placemark.io tool, I deleted the other 8 lines but kepts the 2 polygons. Then in File > Export, I downloaded the file and renamed it 2streets.geojson.
-
Now start up https://Mapshaper.org to upload the crashes.csv (points) and the 2streets.geojson (polygons) in order to join them. See my more detailed instructions about joining at https://github.com/bikewesthartford/wh-crash-analysis
-
In the image below, see how I uploaded the crashes.csv (in this repo), then the 2streets.geojson (in this repo), then joined them. The tool describes how 2 source records (Farmington Ave and Main St) were joined to 2,082 crashes, while 21,122 received no data because they are not located inside any polygons. When I Inspect crashes with my cursor, notice how some of them now contain "Main St" or "Farmington Ave" in a new corridor field. Finally, notice how the polygon outlines appear in light green because I intentionally made them visible (by clicking the eyeball) in the Mapshaper 2streets layer.
- In Mapshaper, go to Export and select only the crashes layer, to be exported as a CSV spreadsheeet. Rename the downloaded file to something like crashes-joined.csv because it now has more info than the original crashes.csv file. See the file with this name in this repo. Open it up and you'll see how you can create a pivot table to show the total number (and percent) of all crashes that happened inside your polygons named Farmington Ave and Main St.
