Preferred methods for Geocoding addresses?

Hey Jeremy!

You are correct, Tableau does not allow you to visualize street address info. That said, if you convert those into Lat-Log points (2 columns), Tableau can accurately populate them on a map. 

I've previously used ArcGIS to geocode street address data. We have concurrent licenses in my department which is a potentially cost-saving route ya'll may be able to take. :)

Hello Jeremy,

I have used ArcGIS and SAS to geocode address data for years. Previously I had used SAS to clean the data into variables for each address segments (street, city, state, zip code) as well as a single variable with all address information and geocoded in ArcGIS. 

I also downloaded the maps and look up tables from SAS's online website and geocoded entirely within SAS. SAS then produced a dataset with the latitude and longitude in decimal degrees I can't remember the projection it uses but be mindful of that, the standard project used is NAD 1983 I believe. 

While I have not done this in R or python I imagine similar resources are available and the process would be similar to SAS. Especially since R and Python are free. Also you can run SAS scripts in python, and I think R, so you can definitely utilize the maps and look up table resources online.

Good luck with your project!
~Maria

Hi Jeremy, I've done geocoding in R before and there's several ways and several packages it can be done with.  There's an R package called tidygeocoder that will connect to several different api's for selection of the data. There's also several examples online using ggmap package with Google's api. The OpenStreetMap api is a good one to use, and there are others like HERE and OpenCage api's also besides the big name ones (Esri, Bing, etc) but some may have limits etc like 2500 at a time. Another way is that if you are aren't comfortable with the coding aspects, QGIS software is a free and open source gis software comparable to ArcGIS that has several geocoding packages in it that are available with more point/click interface. I have used that for more stand alone one-off type projects when needed to geocode points, and may be the easiest approach, outside of the ArcGIS approach that Miranda mentioned, but if you need reproducibility and something that will be continual used over time with new points continually coming in, then R is probably the best approach for that. Hope this helps. 

tidygeocoder package help description
https://cran.r-project.org/web/packages/tidygeocoder/tidygeocoder.pdf 

example 
https://engineerpaige.com/geocoding-in-r/

Thank you all so much for the suggestions and advice! While ArcGIS might've been a possibility, previous attempts at obtaining licenses have not necessarily gone over very fruitfully, so I was hesitant to try and go down that rabbit hole again...

After doing some exploring, Jason 's suggestion tidygeocoder suggestion ended up working very well for me! I only wish I had started with trying that out haha. Thanks again everyone, but totally down to hear what other folks do if there are more methods out there! Hope your projects are all going well!

This is an example of a markdown file that you can use to access python and also run python code from R studio. Please let me know if you want to run through this. 

```{r COVID Geocoding}

 

library(reticulate)   # Bridging R and Python

library(dplyr)        # Data wrangling

library(sf)           # Working with spatial objects

 

save_shp <- paste0("D:\\COVID\\Output\\Geocode_vac_geocode.shp")

 

save_vax <- paste0("Geocode_vaxx.csv")

 

use_python("C:/Program Files/ArcGIS/Pro/bin/Python/envs/arcgispro-py3/python.exe", required = TRUE)

arcpy <- import("arcpy")

 

arcpy$env$overwriteOutput =TRUE

 

arcpy$GeocodeAddresses_geocoding(

  in_table = save_vax,

  address_locator = "//phs110.acphd.ph.local/DEPT/Cape\\GIS\\Composite.loc",

  in_address_fields = "'Street or Intersection' Address VISIBLE NONE;'City or Placename' RECIP_ADDRESS_CITY VISIBLE NONE;State RECIP_ADDRESS_STATE VISIBLE NONE;'ZIP Code' RECIP_ADDRESS_ZIP VISIBLE NONE",

  out_feature_class = save_shp)

 

save_folder <- paste0("D:\\COVID\\Output\\")

#X:\nCoV2020\Data\COVID-19_Analyses\R Scripts\Covid VM

 

```

 

```{python}

import arcpy, sys, string, os, traceback, re, time #, datetime

from arcpy import env

 

for attempt_no in range(5):

  try:

    arcpy.env.overwriteOutput = True

    inpoints = r.save_shp

    outfolder = r.save_folder

    outname = "VaxJoins"

    outpath = outfolder + os.sep

    temp01 = r"in_memory" + \\temp01

    temp02 = r"in_memory" + \\temp02

    temp03 = r"in_memory" + \\temp03

    arcpy.AddXY_management(inpoints)

    arcpy.CalculateField_management(inpoints, "X", '!POINT_X!', "PYTHON_9.3")

    arcpy.CalculateField_management(inpoints, "Y", '!POINT_Y!', "PYTHON_9.3")

    arcpy.SpatialJoin_analysis (inpoints, r"D:\\COVID\\Geocoding\\Places incl water.shp", temp01)

    arcpy.AddMessage("Places")

    arcpy.SpatialJoin_analysis (temp01, r"D:\\COVID\\Geocoding\\CT10 incl water.shp", temp02)

    arcpy.AddMessage("CT10")

    arcpy.SpatialJoin_analysis (temp02, r"D:\\COVID\\Geocoding\\AC vs US.shp", temp03)

    arcpy.AddMessage("County")

  except ValueError as error:

    if attempt_no < (num_retries - 1):

      print("Error: Retrying")

    else:

      raise error

 

with arcpy.da.UpdateCursor(temp03,["Postal","Match_addr"]) as cursor:

  for row in cursor:

    if (row[0] == ' ' and row[1] != ' '):

      row[0] = row[1][-5:]

    cursor.updateRow(row)

with arcpy.da.UpdateCursor(temp03,["Postal","Match_addr"]) as cursor:

  for row in cursor:

    if (row[0][-5:-4] == '-'):

      arcpy.AddMessage(row[0][-5:-4])

      row[0] = row[1][-10:-5]

      arcpy.AddMessage(row[1][-10:-5])

    cursor.updateRow(row)

 

arcpy.AlterField_management(temp03, 'USER_RECIP', 'RECIP_ID')

arcpy.AlterField_management(temp03, 'USER_REC_1', 'ADDRESS1')

arcpy.AlterField_management(temp03, 'USER_REC_2', 'ADDRESS2')

arcpy.AlterField_management(temp03, 'USER_REC_3', 'County_geo')

arcpy.AlterField_management(temp03, 'USER_REC_4', 'zipcode')

arcpy.AlterField_management(temp03, 'USER_REC_5', 'state_geo')

arcpy.AlterField_management(temp03, 'USER_ADDRE', 'geo_address')

arcpy.AlterField_management(temp03, 'IN_Street', 'Address')

arcpy.AlterField_management(temp03, 'USER_full_', 'full_address')

 

fieldNeedList = ["Match_addr",  "POINT_X", "POINT_Y", "NAME", "Place", "CT10", "Postal", "RECIP_ID", "ADDRESS1", "ADDRESS2", "County_geo", "zipcode", "state_geo", "geo_address","Address","full_address"]

 

 

fieldNameList = []

fieldObjList = arcpy.ListFields(temp03)

for field in fieldObjList:

  if not field.required:

    if not field.name in fieldNeedList:

      fieldNameList.append(field.name)

        

arcpy.DeleteField_management(temp03, fieldNameList)

arcpy.AlterField_management(temp03, 'POINT_X', 'X')

arcpy.AlterField_management(temp03, 'POINT_Y', 'Y')

arcpy.AlterField_management(temp03, 'NAME', 'County')

arcpy.AlterField_management(temp03, 'Postal', 'Zip')

arcpy.CopyFeatures_management(temp03, outpath + outname + ".shp")

arcpy.Delete_management("in_memory") 

 

```