Question

我正在开展一个公共卫生项目，我正在为整个烟草商店刮刮整个州。我一直在尝试使用google places API网络服务，特别是Nearby and Text methods。下面是一些快速的R代码（需要一些错误检查，但运行），以向您展示我正在做的两个版本。

我发现关键字搜索的行为似乎与名称搜索相似 - 我的结果是＆＃34; vape＆＃34;在附近的搜索请求或文本搜索请求过于严格，只给我，例如＆＃34; Vape Escape＆＃34;，＆＃34; Cloud Vape Lounge＆＃34;和＆＃34; Juice Vapeorium＆＃34; - 但是当我手动搜索谷歌地图中的相同位置并搜索"vape"或"vape near me"时，我会得到十几个更好的结果，包括＆＃34; Vapor girl＆＃34;和＆＃34; Hazmat＆＃34;没有＆＃34; vape＆＃34;在名称中。

任何想法我做错了，API似乎是将我的关键字请求解释为名称请求？

library(RCurl)
library(RJSONIO)

# NOTE: Radar search may be better here, eventually. https://developers.google.com/places/web-service/search
place_url = function(lat, lon, rad = 5000, kw, return.call = "json", sensor = "false") {
  root = "https://maps.googleapis.com/maps/api/place/nearbysearch/"
  api_key = "MY KEY WENT HERE"
  u <- paste(root, return.call, "?location=", lat, ",", lon, "&radius=", rad, "&keyword=", kw, "&key=", api_key, sep = "")
  return(URLencode(u))
}

text_url = function(lat, lon, rad = 5000, kw, return.call = "json", sensor = "false") {
  #10X multiplier! Be careful!
  root = "https://maps.googleapis.com/maps/api/place/textsearch/"
  api_key = "MY KEY WENT HERE"
  u <- paste(root, return.call, "?location=", lat, ",", lon, "&radius=", rad, "&query=", kw, "&key=", api_key, sep = "")
  return(URLencode(u))
}


place_scrape <- function(keyword, lat, lon, radius=35405.6, verbose=TRUE) {
  # radius=35405.6: ~=22 miles. circular radius means ~11x11 mile square w/ overlap. Treat like 10x10, dump dups.
  df = data.frame(name=character(), lat=double(), lon=double(), id=character(), vicinity=character()) #for starters

  if(verbose) cat("Looking for '", keyword,"' around lat", lat, "and lon", lon, "\n")
  #lat = 35.917217; lon = -79.054554; radius = 35405.6; keyword = "vape near me"; verbose=T
  u=place_url(lat,lon, radius, keyword) 
  #u=text_url(lat,lon, radius, keyword) 
  doc <- getURL(u)
  x <- fromJSON(doc,simplify = FALSE)
  length(x)
  if(verbose) cat("Found", length(x), "results.\n")

  for (i in 1:length(x)){
    result = data.frame(name=x$results[[i]]$name, 
                        lat= x$results[[i]]$geometry$location$lat, lon = x$results[[i]]$geometry$location$lng, 
                        id = x$results[[i]]$id, 
  #                      vicinity=x$results[[i]]$formatted_address) #text version
                        vicinity=x$results[[i]]$vicinity) #place version
    df = rbind(df, result)
  }
  if(verbose) print(df)
  return(df)
}

# 35.917217, -79.054554 is chapel hill
results = place_scrape("vape", 35.917217, -79.054554)
results = place_scrape("vapor", 35.917217, -79.054554)
results = place_scrape("vapes", 35.917217, -79.054554)


#scrape maryland: from 40N to 38.5. ~360 miles across by

为什么关键字搜索会像Google地方API中的名称搜索一样？

0 个答案: