Read Me for Zillow Project

Codes

(0) initialize_zillow.do
* This .do file does the following:
*	- makes a master list of property IDs with addresses etc
* 	- stacks all the property attibute data and then merges the master property list

(0.1) zillow_property_list.do
* this program is used to make a unique property list
* this is done by doing the following...
*	(1) removing duplicate properties
* 	(2) removing any spaces and extra symbols in "housenumber" and "streetsuffix"
*	(3) isolating problematic properties. often there are non numbers in the numeric categories.
*	(4) broaden criterion and refine again

(1) zillow_trimdown.do
* this code times down the zillow data to useable observations
* 	(1) merge data with analysis table. we then remove unneeded variables. 
*	(2) split data into numeric and string values
*	(3) take the numeric data set and use it to tag unlikely single-family homes (SFH)

(2) zillow_76_to_text.do
* this code does a quick clean of atype 76
* 	(1) keep only atype 76. remove all non alpha-numeric characters. replaces with spaces.
*	(2) export text file

(3) run_zillow_ngrams.py
* simple program that calls the function "zillow_ngrams.py"
* depends on:
	- zillow_ngrams.py
		main workhorse program. tokenizes every word (uni, bi, and tri grams) in the text file created by zillow_76_to_text.do
	- replace_spanish.py
		replaces all spanish characters with english versions
	- remove_html.py
		removes html formatting
	- nolla_lang_detect.py
		inspired by a program written by a dude last named "Nolla". it scans text and detects if its english, spanish, or bi lingual.
	
(4) zillow_logit.do
*	imports data, trims counts, then runs logit on some bigrams