hannesdatta

/*

===================================
ROLLING WINDOW AGGREGATION IN STATA
===================================

Problem:
--------

Assume you have a data set with time/datestamps, noting

hannesdatta

#!/usr/bin/python

import argparse
from pymongo import MongoClient as Client
from bson import BSON
from bson import json_util
import json
import os

# mongo client

hannesdatta

psignstars <- function(x) {
  sapply(x, function(p) ifelse(p < .01, "***", ifelse(p < .05, "**", ifelse(p < .1, "*", "    "))))
}


# Function to run regression model (#mmix, with spec formula)
regmodel <- function(formula=list(~1+I(country_class=='linc') + as.factor(category) + as.factor(brand)), 
                     dat, model = 'lm') {
 
  lmerctrl = lmerControl(optimizer ="Nelder_Mead", check.conv.singular="ignore")

hannesdatta

################################################################
# DOWNLOAD ENTIRE FOLDER STRUCTURE FROM DROPBOX TO LOCAL DRIVE #
################################################################

# Instructions:
# (1) install dropbox API using pip
# > pip install dropbox

# (2) Create application to make requests to the Dropbox API
# - Go to: https://dropbox.com/developers/apps

hannesdatta

# PROBLEM:

# I would like to give the new column in the new DT 
# the name "mean_price"; however, I cannot figure out how to do this.

# It should be possible but I don't know how. 
# Here is someone with a related issue: https://stackoverflow.com/questions/12391950/select-assign-to-data-table-when-variable-names-are-stored-in-a-character-vect
# Do you know how to resolve this issue?

# EXAMPLE:

hannesdatta

This gist has been replaced by an R package with an updated list of labels.

Get it on GitHub: https://github.com/hannesdatta/musicMetadata

LEGACY CODE

#################################################
#                                               #
# Classify music labels                         #
# into major labels (Sony, Warner, Universal),  #

hannesdatta

####################################
#                                  #
# UNIT ROOT TESTS                  #
# IN THE ABSENCE OF                #
# KNOWLEDGE ON THE ACTUAL          #
# DATA GENERATING PROCESS          #
#                                  #
# Enders 1995,                     #
# Applied Econometric Time Series  #
# pp. 254 - 258 and                #

hannesdatta

require(stringi)

spelling_variants <- function(x, remove_collabs=F, remove_parentheses=T) {
  qualifiers = c(" feat .*", " feat[.].*", " ft.*", " ft[.].*"," featuring.*"," vs[.].*"," vs.*"," versus.*"," with.*","[-].*"," / .*",
                 "/.*","[|].*", "[[].*[]]", "[)].*", ";.*","[+].*","[&] .*","[&].*",",.*"," and .*", " con .*", " e .*", " et .*",
                 " x .*")
  
  # remove articles (a, the)
  ret = gsub(" a ", "", tolower(str_trim(x)))
  

hannesdatta

# TRANSFORMS A SET OF URLS TO TINY-URLS (SHORTENED URLS)
# adapted from https://www.geeksforgeeks.org/python-url-shortener-using-tinyurl-api/

hannesdatta

# Quick conversion of `character` date columns to Date format using data.table
# fread(..., colClasses = c(date='Date')) is slow for large data sets, especially when
# the number of unique dates is small, but the number of cross-sectional units is large.

# The intuition of this algorithm is to only convert the UNIQUE dates to dates using as.Date,
# and then merging them back to the original data.table.

library(data.table)

data.table.date <- function(dt, datecol) {