WalkerHarrison

import json
import os
import datetime
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from collections import Counter

files= os.listdir('Searches')
del files[0]

WalkerHarrison

hours = [datetime.datetime.strptime(i, '%Y-%m-%d %H:%M:%S').hour for i in dates]
n, bins, patches = plt.hist(hours, 24, facecolor='blue', alpha=0.75)
plt.xticks([0,6,12,18], ['12 AM','6 AM', '12 PM', '6 PM'], fontsize=18)
plt.xlabel('Hour', fontsize=24)
plt.ylabel('Frequency', fontsize=24)
plt.gcf().set_size_inches(18.5, 10.5, forward=True)
plt.show()

WalkerHarrison

combo = ' '.join(searches)
freqs = Counter(combo.split())
top = freqs.most_common(40)

words = []
counts = []
for i in range(40):
    words.append(top[i][0])
    counts.append(top[i][1])

WalkerHarrison

d = {"search": searches,
    "time": dates}
googled = pd.DataFrame(d)

dt = datetime.datetime(2014, 10, 1)
end = datetime.datetime(2017, 3, 5)
step = datetime.timedelta(days=7)

weekly = []

WalkerHarrison

library(XML)
library(ggplot2)
library(ggmap)
library(RSocrata)

url_trees <- 'https://data.cityofnewyork.us/resource/nwxe-4ae8.csv'
url_zips <- 'http://zipatlas.com/us/ny/brooklyn/zip-code-comparison/median-household-income.htm'

trees <-read.socrata(url_trees)
trees <- subset(trees, boroname == 'Brooklyn')

WalkerHarrison

extras <- do.call(data.frame, aggregate(tree_dbh ~ zipcode, trees, 
                                        FUN=function(x) c(mn = mean(x), count = length(x))))
neighborhoods <- merge(neighborhoods, extras, by="zipcode")
neighborhoods$trees_per_capita <- neighborhoods$tree_dbh.count/neighborhoods$population

ggplot(neighborhoods, 
       aes(x=income, y=tree_dbh.mn, size = trees_per_capita, label=neighborhoods$zipcode)) +
  geom_point(color="green") + geom_smooth(method='lm',formula=y~x, show.legend =F) +
  geom_text(size=4, nudge_x = 1300, nudge_y = c(0.1, -0.1, -0.1)) +
  scale_size_continuous(range=c(0,10)) +

WalkerHarrison

zoom <- subset(trees, zipcode %in% c(11239, 11206, 11212, 11224, 11221, 
                                     11201, 11215, 11217, 11231, 11234))
zoom$hood <- as.factor(ifelse(zoom$zipcode %in% c(11201, 11215, 11217, 11231, 11234), 1,0))

map <- get_map(location = c(lon = -73.95, lat = 40.64), zoom = 12,
               maptype = "satellite", source = "google")
ggmap(map) + geom_point(data=zoom, aes(x = longitude, y = latitude, col = hood),
                        size = 0.5, shape = 16, alpha = 0.1, show.legend = F)

WalkerHarrison

zoom$cherry <- grepl("cherry", zoom$spc_common)
zoom$dead <-zoom$status == "Dead"
table(zoom$hood, zoom$dead)
table(zoom$hood, zoom$cherry)
table(zoom$hood, zoom$brch_shoe)

WalkerHarrison

library(ggplot2)
set.seed(1)

x <- 1:100
y <- x^2*sin(2*pi*x/100) + 500*rnorm(length(x))

df <- data.frame(x, y)
h <- 12
smoother <- data.frame(ksmooth(x, y, "normal", bandwidth = h, n.points = 100))

WalkerHarrison

library(pracma)

scale <- abs((erfinv(-0.5)*(2^0.5)*4/h)^-1)
wt <- dnorm(x-50, 0, scale)

ggplot(df, aes(x, y, col = wt)) + 
  geom_point(size = pmax(100*wt, 1)) + 
  geom_line(data = smoother, aes(x, y), col = "black") + 
  geom_point(data = smoother[x==50,], aes(x, y), size = 3,
             col = "black", shape = 21, fill = "white") +