Yelp Yep -- R Shiny Project

Introduction

In this project, the app aims to identify the key features for people in Phoenix to give score on Yelp. Using the Yelp Dataset from Yelp Dataset Challenge, the app compares the influence of some attributes in the dataset based on the category. In particular, the app analysis every attribute to the appearance of hipsters in order to find out whether the ambiance of hipsters will affect the average score of the store.


A brief show case, click to view

Data Set

According to the description of Yelp Challenge, this dataset includes:
4.1M reviews and 947K tips by 1M users for 144K businesses
1.1M business attributes, e.g., hours, parking availability, ambience.
Aggregated check-ins over time for each of the 125K businesses200,000 pictures from the included businesses

It includes 11 cities such as: Edinburgh in U.K., Karlsruhe in Germany, Montreal and Waterloo in Canada, Pittsburgh, Charlotte, Urbana-Champaign, Phoenix, Las Vegas, Madison, Cleveland in U.S.
The whole dataset is composed by five json files: business, checkin, review, tip, and user file. By using ndjson
package, we read data as a data frame. The app only joins the business and review files by business_id
in order to get the all business attributes, review counts, star rating.


data sample


Data summary

Exploratory Data Analysis

From the graph of category and total reviews, we can see that restaurant's Yelp strongly influences an individual's dining decisions. So the app mainly focus on the stores provide foods.

The Yelp Dataset comes from the Yelp Dataset Challenge webpage. Our project only focus on the Phoenix so we filtered out other countries and states. This left us with 10,629 businesses. I inner joined the business and review table so I have 10,629 observations and 116 variables. In order to directly find out the type of store has the most ambience of hipsters, I also filtered out all non-restaurant business and build a subset dataset restaurant.

Screen Shot 2017-02-04 at 9.39.33 PM

From plotting, I find out that there are some features corresponding to high star rating include: review count, noise level, outdoor seating, classy ambience, hipster ambience, good for kids, good for groups, divey ambience, garage parking, and has TV. For most of the plotting, the app shows that stores that provide food have more count and the average stars in the range from 3.0 to 4.5.

One of the interesting things from the data is that for stores that good for kids category, Hair Salons and Active Life shows a outstanding high score which makes lots of sense since many parents especially moms would always go to those places and they need to take care of their children at the same time. If those places have good environment for kids such as day care center, it is not difficult to imagine how much work they will save for moms. The same situation also happens for the Garage Parking in Hair Salons. The data shows that the hair salon stores provide garage parking will always receive higher scores. Those results also let us start to think that the business area which provide service can get a higher score if they keep making customers more convenient by providing parking plot and better environment for kids, etc.

Who are hipsters?

According to the Google translator, a hipster is:

a person who follows the latest trends and fashions, especially those regarded as being outside the cultural mainstream.

Hipsters are those people who walk around town as a beard-and-glasses with plaid shirts, listening to new-ish music and seeking status. For some reasons, many people hates hipsters. The Yelp data I have also take this into consideration when rating a store. I specifically do some plotting and try to find out whether the ambience of hipster will influence the rating of the store. It turns out that hipster independently would not affect the lower rating at all. However, one of the interesting thing is that hipster would normally show up in the food, bars, American restaurants. They seldom go to the Asian restaurant except the fusion bars which is much similar with American style bars. Also, I find out hipsters are not the main source of the noise. So in my opinion, it is unreasonable to discriminate against them. Even more, take the ambience of hipster into account of score ranking itself is a discrimination.


Why people hate hipsters?

Quora says that people hipsters for different reasons.

  • The recent movement of hipsterius civilatus (family name) comes from young middle- and upper-class citizens who are creating their own counter-culture movement. The reason for the hate is because they are generally seen as spoiled, have a certain categoric smugness to themselves.
  • Society's perceptions of youth culture (in other news, see: rock and roll, disco, hippie, grunge, yuppie, emo, punk, and so on)
  • Certain key attributes and attitudes that hipsters are seen to have, which include; vegetarianism or veganism, concern about the environment, anti-capitalist, anti-consumerist, a strong love for independent music and movies.

Data Analysis related to hipsters

A interesting result is that hipsters seldom shows up in the Asian restaurant except some Asian Fusion store from data. The data shows that hipsters more into bars, Gastropubs, American food, Mexican food, Pizza, Sandwiches, Burgers, Art & Entertainment.

The app also can let us pick two attributes to see the relationship between two attributes. From the Mosaic Plot, the app shows that places provide the outdoor seating will have more chance of ambience of hipsters.

A strange finding is that the histogram shows that hipsters are fond of the place that good for groups, however, it does not show anything from mosaic plotting. A counter example is noise level and the ambience of hipster, both bar charts and mosaic plot shows that there is not any direct relationship between them. I cannot explain now why the result from mosaic plot and bar charts is different. But I am sure there must be some statistic insights about how we evaluate the relationship between these two attributes in different type of plotting. Moreover, mosaic plot tells us that hipsters like to hangout in the place with the price range from 1 to 3. Also, the data shows that places provide outdoor seating are not good for kids at the same time.

The app specifically compares the ambience of hipsters and ratio of other business attributes such as food good for group, noise level, good for kids, outdoor seating, credit card usage, divey, garage parking, has TV, price range, take out option, reviews count. Regard to the noise level, one of the interesting thing is that hipsters has nothing to do with the noise level which is the opposite to the most people's expectation. For example, as for Arts & Entertainment category, we can see that if there is hipsters the noise level is high but if there is not any ambience of hipsters, the noise level is higher than without hipsters show up. Another example is that Seafood category, we can see that without hipsters, Seafood store is much quieter than the ambience of hipsters. So we cannot sure that hipsters will bring up the noise level precisely. In addition, hipsters seems like stores with outdoor seating. They also seems into divey and place with garage parking.

Maps

The app shows two maps. First of all, the map marks the place where has the ambience of hipsters. The pop up can tell you the place's name, address and score. If you really are not a fan of hipsters, you can avoid them. Or if you like me who really do not mind, you can use it as a ranking reference. Besides, if you are a potential hipsters or big fan of hipsters, welcome to go to those place and make friends with them.

The second map is a density map built on the ambience of hipsters. From it you can tell the rough area of the hipsters. The redest area is where hipsters have a larger chance to hang out.

Future work
In the future, I plan to build a social network between all users and their friends so that I can build a small recommendation system based on their common tastes.

最后編輯于
?著作權歸作者所有,轉載或內容合作請聯系作者
平臺聲明:文章內容(如有圖片或視頻亦包括在內)由作者上傳并發布,文章內容僅代表作者本人觀點,簡書系信息發布平臺,僅提供信息存儲服務。
  • 序言:七十年代末,一起剝皮案震驚了整個濱河市,隨后出現的幾起案子,更是在濱河造成了極大的恐慌,老刑警劉巖,帶你破解...
    沈念sama閱讀 230,321評論 6 543
  • 序言:濱河連續發生了三起死亡事件,死亡現場離奇詭異,居然都是意外死亡,警方通過查閱死者的電腦和手機,發現死者居然都...
    沈念sama閱讀 99,559評論 3 429
  • 文/潘曉璐 我一進店門,熙熙樓的掌柜王于貴愁眉苦臉地迎上來,“玉大人,你說我怎么就攤上這事。” “怎么了?”我有些...
    開封第一講書人閱讀 178,442評論 0 383
  • 文/不壞的土叔 我叫張陵,是天一觀的道長。 經常有香客問我,道長,這世上最難降的妖魔是什么? 我笑而不...
    開封第一講書人閱讀 63,835評論 1 317
  • 正文 為了忘掉前任,我火速辦了婚禮,結果婚禮上,老公的妹妹穿的比我還像新娘。我一直安慰自己,他們只是感情好,可當我...
    茶點故事閱讀 72,581評論 6 412
  • 文/花漫 我一把揭開白布。 她就那樣靜靜地躺著,像睡著了一般。 火紅的嫁衣襯著肌膚如雪。 梳的紋絲不亂的頭發上,一...
    開封第一講書人閱讀 55,922評論 1 328
  • 那天,我揣著相機與錄音,去河邊找鬼。 笑死,一個胖子當著我的面吹牛,可吹牛的內容都是我干的。 我是一名探鬼主播,決...
    沈念sama閱讀 43,931評論 3 447
  • 文/蒼蘭香墨 我猛地睜開眼,長吁一口氣:“原來是場噩夢啊……” “哼!你這毒婦竟也來了?” 一聲冷哼從身側響起,我...
    開封第一講書人閱讀 43,096評論 0 290
  • 序言:老撾萬榮一對情侶失蹤,失蹤者是張志新(化名)和其女友劉穎,沒想到半個月后,有當地人在樹林里發現了一具尸體,經...
    沈念sama閱讀 49,639評論 1 336
  • 正文 獨居荒郊野嶺守林人離奇死亡,尸身上長有42處帶血的膿包…… 初始之章·張勛 以下內容為張勛視角 年9月15日...
    茶點故事閱讀 41,374評論 3 358
  • 正文 我和宋清朗相戀三年,在試婚紗的時候發現自己被綠了。 大學時的朋友給我發了我未婚夫和他白月光在一起吃飯的照片。...
    茶點故事閱讀 43,591評論 1 374
  • 序言:一個原本活蹦亂跳的男人離奇死亡,死狀恐怖,靈堂內的尸體忽然破棺而出,到底是詐尸還是另有隱情,我是刑警寧澤,帶...
    沈念sama閱讀 39,104評論 5 364
  • 正文 年R本政府宣布,位于F島的核電站,受9級特大地震影響,放射性物質發生泄漏。R本人自食惡果不足惜,卻給世界環境...
    茶點故事閱讀 44,789評論 3 349
  • 文/蒙蒙 一、第九天 我趴在偏房一處隱蔽的房頂上張望。 院中可真熱鬧,春花似錦、人聲如沸。這莊子的主人今日做“春日...
    開封第一講書人閱讀 35,196評論 0 28
  • 文/蒼蘭香墨 我抬頭看了看天上的太陽。三九已至,卻和暖如春,著一層夾襖步出監牢的瞬間,已是汗流浹背。 一陣腳步聲響...
    開封第一講書人閱讀 36,524評論 1 295
  • 我被黑心中介騙來泰國打工, 沒想到剛下飛機就差點兒被人妖公主榨干…… 1. 我叫王不留,地道東北人。 一個月前我還...
    沈念sama閱讀 52,322評論 3 400
  • 正文 我出身青樓,卻偏偏與公主長得像,于是被迫代替她去往敵國和親。 傳聞我的和親對象是個殘疾皇子,可洞房花燭夜當晚...
    茶點故事閱讀 48,554評論 2 379

推薦閱讀更多精彩內容

  • 感恩讀書會的邀請,前往參加 感恩死黨們組辦的讀書會,參與瑜伽的學習 感恩金華的教學,舒服的感覺! 感恩朋友的邀請,...
    我不叫許仲斌閱讀 257評論 0 3
  • 我有個設計公司的采購,以前她是做汽車配件的,朋友介紹就認識我了,后來給我做了一件貨,但交貨期方面,我推了兩次,因為...
    六爺有話說閱讀 718評論 0 0
  • 澤川驅車將于行, 忽聞江油裂地聲。 揮袖辭別長央宮, 長越千里過漢城。 萬壑千巖等閑過, 亂云叢中穿霧行。 渝鄂巫...
    禹業立閱讀 138評論 0 0