kfc流程管理炸薯?xiàng)l幾秒

In February, our Data Science team had an argument about which restaurant we went to made the best French Fry.

2月，我們的數(shù)據(jù)科學(xué)團(tuán)隊(duì)對(duì)我們?nèi)ツ募也蛷d做得最好的炸薯?xiàng)l產(chǎn)生了爭(zhēng)議。

We decided to make it a competition throughout Feb/March and find out who truly was King of the French Fry. Move over March Madness, Hello Starch Madness

我們決定在整個(gè)2月/ 3月進(jìn)行比賽，以找出誰(shuí)是真正的炸薯?xiàng)l之王。 移動(dòng)三月瘋狂 ，你好淀粉瘋狂

I typed up our findings into the first academic paper about French Fries and submitted it to arXiv.

我將我們的發(fā)現(xiàn)輸入關(guān)于薯?xiàng)l的第一篇學(xué)術(shù)論文中 ，并將其提交給arXiv。

It was promptly rejected. :(

它很快被拒絕了。 :(

Nevertheless, here is the story of The Objectively Best French Fries, or as I titled it:

盡管如此，這還是The客觀上最好的炸薯?xiàng)l的故事，或者如我所說(shuō)：

An Objective Analysis of the Subjective Quality of French Fries in the Downtown Santa Monica Area

市中心圣莫尼卡地區(qū)薯?xiàng)l主觀質(zhì)量的客觀分析

Warning: This article will very likely make you hungry for French Fries

警告 ：這篇文章很可能會(huì)讓您餓到炸薯?xiàng)l

抽象 (Abstract)

It all started with a team lunch. We always ordered fries for the table. After several lunches, we began to ask the question: What are the best French fries near our Santa Monica office? How do you even determine “best”? As a Data Science company, we knew this couldn’t be determined simply by eating fries at various restaurants — we had to structure our analysis around scientific rigor and specified constraints, so that we could be confident in our findings… If the results weren’t reproducible through peer review, what was the point? And thus, Starch Madness was born; a contest wherein a bunch of data nerds objectively decided which restaurant served the subjectively best French fries in Santa Monica.

這 一切都始于一個(gè)團(tuán)隊(duì)午餐。 我們總是為桌上點(diǎn)薯?xiàng)l。 幾頓午餐后，我們開(kāi)始提出一個(gè)問(wèn)題：圣莫尼卡辦公室附近最好的炸薯?xiàng)l是什么？ 您甚至如何確定“最佳”？ 作為一家數(shù)據(jù)科學(xué)公司，我們知道不能僅僅通過(guò)在各家餐館吃薯?xiàng)l來(lái)確定這一點(diǎn)-我們必須圍繞科學(xué)的嚴(yán)格性和特定的約束來(lái)構(gòu)建分析，以便我們對(duì)我們的發(fā)現(xiàn)充滿信心……如果結(jié)果不是'通過(guò)同行評(píng)審可重現(xiàn)，意義何在？ 因此，瘋狂淀粉誕生了。 一場(chǎng)比賽，一群數(shù)據(jù)書(shū)呆子客觀地決定了哪家餐廳在圣莫尼卡提供主觀上最好的炸薯?xiàng)l。

介紹 (Introduction)

In 1802, Thomas Jefferson served “potatoes served in the French manner” at a White House dinner. From there, the French fry was born. Potatoes cut into slices and deep fried became a staple in American cuisine.

1802年，托馬斯·杰斐遜擔(dān)任“法國(guó)送達(dá)方式土豆”在白宮晚宴。 炸薯?xiàng)l從那里誕生。 土豆切成薄片然后油炸成為美國(guó)菜的主食。

Despite being most associated with burgers and fast food, French fries are served in most restaurants. With their ubiquity, the question must be asked: “Who serves the best French fry?” The simplicity of the food recipe makes this question especially interesting, as there are only a few variables a chef can play with to craft their ideal French fry.

盡管炸薯?xiàng)l與漢堡和??快餐最相關(guān)，但大多數(shù)餐館還是供應(yīng)炸薯?xiàng)l。 由于它們無(wú)處不在，所以必須提出一個(gè)問(wèn)題：“誰(shuí)為法國(guó)炸薯?xiàng)l做的最好？” 食物食譜的簡(jiǎn)單性使這個(gè)問(wèn)題變得尤為有趣，因?yàn)閺N師只能選擇幾個(gè)變量來(lái)制作其理想的炸薯?xiàng)l。

As a California Native, In-n-Out Fries hold a special place in my heart — Photo by In-N-Out Inc.作為加利福尼亞州的原住民，薯?xiàng)l在我的心中占有一個(gè)特殊的地方—照片作者In-N-Out Inc.

In addition, some institutions produce mass quantities of fries per day — does this mass production reduce the quality of the fry? Does the fact that they must focus a large portion of their business on producing enjoyable fries create a better quality than a restaurant known for higher end cuisine?

另外，一些機(jī)構(gòu)每天生產(chǎn)大量的薯?xiàng)l-這種大量生產(chǎn)會(huì)降低魚(yú)苗的質(zhì)量嗎？ 他們必須將大部分精力都集中在生產(chǎn)美味的薯?xiàng)l上，這是否比以高端美食聞名的餐廳產(chǎn)生了更好的質(zhì)量？

We would like to take this moment to address the research elephant in the room — if this research is even important in the first place. We propose that in fact:

我們想借此機(jī)會(huì)談?wù)劮块g內(nèi)的研究大象-如果這項(xiàng)研究首先很重要。 我們實(shí)際上建議：

是的。 (Yes it is.)

The pursuit of knowledge is admirable regardless of scale, and there has yet to be any studies conducted on the French fry quality of the downtown Santa Monica area. We may not have millions of dollars in French fry funding (yet) at our disposal, but we have the tenacity and drive (and palettes) to see this project through. It is our pleasure to add a tiny morsel of new knowledge into the human corpus.

無(wú)論規(guī)模大小，對(duì)知識(shí)的追求都是令人欽佩的，并且尚未對(duì)圣莫尼卡市區(qū)的炸薯?xiàng)l質(zhì)量進(jìn)行任何研究。 我們可能尚未獲得數(shù)百萬(wàn)美元的炸薯?xiàng)l資金，但我們有足夠的毅力和動(dòng)力(和調(diào)色板)來(lái)完成該項(xiàng)目。 我們很高興在人類(lèi)語(yǔ)料庫(kù)中添加一小撮新知識(shí)。

1.方法 (1. Method)

1.1 Location

1.1位置

We had to first choose the fries that would be entered into the contest. The simplest sampling method was to select fries that could be purchased at a location that was less than 10min walk away from our office. Luckily, our office was located in the heart of downtown Santa Monica and there was no shortage of fry options; in fact there were far too many. To narrow down the locations, we simply typed “best French fries” into Yelp and chose the top 12 results. Several of the locations were eventually replaced due to fry style (see section 1.2) but this methodology yielded a strong crop of contenders.

我們必須首先選擇要參加比賽的炸薯?xiàng)l。 最簡(jiǎn)單的采樣方法是選擇可以在距我們辦公室步行不到10分鐘的位置購(gòu)買(mǎi)的薯?xiàng)l。 幸運(yùn)的是，我們的辦公室位于圣莫尼卡市區(qū)的中心地帶，這里不乏炸魚(yú)的選擇。 實(shí)際上有太多。 為了縮小位置，我們只需在Yelp中鍵入“最佳薯?xiàng)l”，然后選擇前12個(gè)結(jié)果即可。 最終，由于炸魚(yú)的風(fēng)格，一些地點(diǎn)被更換了(見(jiàn)1.2節(jié))，但是這種方法產(chǎn)生了大量的競(jìng)爭(zhēng)者。

To avoid bias from judges based on where the French fries were from, each location was randomly assigned a false identity based on popular T.V shows. Figure 1 is the key mapping the real restaurants to their false identity. Henceforth all restaurants will be referred to by their false identity names.

為了避免法官根據(jù)炸薯?xiàng)l的來(lái)源而產(chǎn)生偏見(jiàn)，根據(jù)流行的電視節(jié)目，每個(gè)地點(diǎn)都被隨機(jī)分配了一個(gè)虛假的身份。 圖1是將真實(shí)餐廳映射到其虛假身份的密鑰。 從此以后，所有餐廳都將使用其虛假的身份名稱來(lái)引用。

Figure 1: Table of False Identities given to each contestant to remove location bias圖1：為消除位置偏見(jiàn)而提供給每個(gè)參賽者的虛假身份表

1.2 Style & Procurement

1.2樣式與采購(gòu)

There are many types of French fries, and unfortunately comparing different styles of French fries is like comparing apples and oranges, and thus we had to standardize our French fry types. We settled on the most common style — salted potato fries. This excluded popular variants such as waffle fries and sweet potato fries, but still allowed creativity and diversity amongst the participants.

?這里有許多類(lèi)型的炸薯?xiàng)l，不幸的是比較不同風(fēng)格的薯?xiàng)l，就像比較蘋(píng)果和桔子，因此我們必須規(guī)范我們的法式炸薯?xiàng)l類(lèi)型。 我們選擇了最普通的風(fēng)格-咸薯?xiàng)l。 這排除了華夫餅薯?xiàng)l和地瓜薯?xiàng)l等受歡迎的變種，但仍使參與者具有創(chuàng)造力和多樣性。

Emmy Smith on 艾美獎(jiǎng)史密斯上UnsplashUnsplash

Judging one batch of fries at a time was not a feasible option, as that would have required quadruple the amount of time the current experiment was allotted as well as made tournament play (1.3.2) impossible. This increase in time would have inevitably resulted in judging fatigue. We instead attempted to gather as many fries as required per round at one time without degrading the integrity of each fry batch.

一次判斷一批炸薯?xiàng)l是不可行的選擇，因?yàn)檫@將需要分配當(dāng)前實(shí)驗(yàn)時(shí)間的四倍，并且使比賽無(wú)法進(jìn)行(1.3.2)。 時(shí)間的增加將不可避免地導(dǎo)致判斷疲勞。 相反，我們嘗試一次收集每輪所需數(shù)量的炸薯?xiàng)l，而不會(huì)降低每批炸薯?xiàng)l的完整性。

Despite our earnest efforts to procure the fry batches as close to simultaneously as possible, our retrieval was not without flaws and some may argue the distances a fry batch travelled correlates inversely with the batch’s score. This conjecture is exhaustively reviewed in our discussion section 3.2.

盡管我們竭盡全力在盡可能近的時(shí)間內(nèi)購(gòu)買(mǎi)炸薯?xiàng)l，但我們的檢索并非沒(méi)有缺陷，有些人可能會(huì)爭(zhēng)辯說(shuō)，炸薯?xiàng)l的行進(jìn)距離與配料的分?jǐn)?shù)成反比。 在我們的討論部分3.2中，對(duì)該推測(cè)進(jìn)行了詳盡的回顧。

Individuals in charge of picking up the French fries were instructed to order the fries by saying, “can I please have an order of fries.” Whatever style the restaurant produced from this order was considered their typical fry style.

負(fù)責(zé)接管炸薯?xiàng)l的人員被指示要點(diǎn)炸薯?xiàng)l，說(shuō)：“請(qǐng)給我一份炸薯?xiàng)l。” 根據(jù)此訂單制作的餐廳無(wú)論哪種風(fēng)格，都被視為其典型的油炸風(fēng)格。

1.3 Experimental Design

1.3實(shí)驗(yàn)設(shè)計(jì)

Similar to the FIFA World Cup tournament design, Starch Madness was structured with two sections: group play (1.3.1) and tournament play (1.3.2). Adequate performance in the group play stage qualified you for tournament play.

小號(hào) imilar國(guó)際足聯(lián)世界杯賽的設(shè)計(jì)，淀粉瘋狂與兩個(gè)部分構(gòu)成：小組賽(1.3.1)和錦標(biāo)賽(1.3.2)。 團(tuán)體賽階段的充分表現(xiàn)使您有資格參加比賽。

Figure 2: Group Play Divisions. “Classic” movie distinction is subjective (but correct)圖2：小組賽分區(qū)。 “經(jīng)典”電影的區(qū)別是主觀的(但正確的)

1.3.1 Group Play Stage

1.3.1小組比賽階段

Each of the 12 participating French fries were randomly assigned into 4 groups of 3 contestants, called divisions, to be rated during the group play stage. The scoring used was the F.A.T System, defined in 1.4. Judges were instructed to use the F.A.T system (1.4) instead of comparing fries from within a division to reach a score.

參加比賽的12條炸薯?xiàng)l中的每一個(gè)都被隨機(jī)分為4組，每組3名選手，稱為部門(mén)，在小組比賽階段進(jìn)行評(píng)分。 使用的評(píng)分是1.4中定義的FAT系統(tǒng)。 指示法官使用FAT系統(tǒng)(1.4)，而不是比較部門(mén)內(nèi)的薯?xiàng)l以取得分?jǐn)?shù)。

This is an important distinction as producing an objective score for each fry is a stronger indicator of fry strength than a comparative one. In addition, if a particular division has three excellent (or terrible) batches, they should be admitted into tournament play on their strength alone, rather than their division’s overall strength. Figure 2 Outlines the divisions.

這是一個(gè)重要的區(qū)別，因?yàn)闉槊總€(gè)油炸物產(chǎn)生一個(gè)客觀評(píng)分是比對(duì)比性更強(qiáng)的油炸強(qiáng)度指標(biāo)。 此外，如果一個(gè)特定的部門(mén)有三個(gè)優(yōu)秀(或糟糕)的批次，則應(yīng)僅憑其實(shí)力而不是其部門(mén)的整體實(shí)力被允許進(jìn)入比賽。 圖2概述了劃分。

1.3.2 Tournament Play

1.3.2比賽比賽

Figure 2: Group Play Divisions. “Classic” movie distinction is subjective (but correct)

圖2：小組賽分區(qū)。 “經(jīng)典”電影的區(qū)別是主觀的(但正確的)

While not the most scientifically rigorous way to compare French fries, it certainly is entertaining. During this stage the top 8 highest rated French fry batches were seeded and placed in a single elimination bracket. Each head-to-head match-up was a blind vote of which fry the judge preferred. Through the tournament play, a winner is crowned.

雖然這不是比較法式炸薯?xiàng)l的最科學(xué)嚴(yán)謹(jǐn)?shù)姆椒?#xff0c;但它確實(shí)很有趣。 在此階段中，對(duì)前8個(gè)收視率最高的炸薯?xiàng)l進(jìn)行播種，并放入單個(gè)淘汰箱中。 每次正面交鋒都是盲目投票，法官傾向于。 通過(guò)比賽，獲勝者加冕。

1.4 The F.A.T System

1.4 FAT系統(tǒng)

One of the first obstacles set forth when judging French fries is how to standardize scoring. And to do this, one must answer a deep philosophical question of, “what makes a good French fry?” Our team of fry-experts narrowed the quality of a French fry down into 3 categories:

第一障礙?NE闡述當(dāng)判斷薯?xiàng)l是如何規(guī)范的得分。 為此，必須回答一個(gè)深刻的哲學(xué)問(wèn)題：“什么能制作出優(yōu)質(zhì)的炸薯?xiàng)l？” 我們的魚(yú)苗專家團(tuán)隊(duì)將法國(guó)魚(yú)苗的質(zhì)量分為以下三類(lèi)：

味道。 (Flavor.)

出現(xiàn)。 (Appearance.)

質(zhì)地。 (Texture.)

This 3-part scoring rubric was brilliantly named the “F.A.T System.” Each of these represents a core attribute of a good French fry. By rating batches based on these metrics, we can decide with confidence which fry is best. Each of these metrics is scored on a 5-point scale, with 1 being the lowest and 5 being the highest. Averaging these three numbers produces a fry’s overall quality and helps the judges make important voting decisions during tournament play.

這個(gè)由三部分組成的評(píng)分規(guī)則被巧妙地命名為“ FAT系統(tǒng)”。 這些都代表了好的炸薯?xiàng)l的核心屬性。 通過(guò)基于這些指標(biāo)對(duì)批次進(jìn)行評(píng)分，我們可以自信地決定哪種魚(yú)苗是最好的。 這些指標(biāo)均以5分制進(jìn)行評(píng)分，其中1分最低，而5分最高。 將這三個(gè)數(shù)字取平均值可得出魚(yú)苗的整體質(zhì)量，并有助于裁判在比賽中做出重要的投票決定。

1.5 Judges

1.5評(píng)委

The French fries were judged by the capable RETINA.ai team, whose love of fries is only surpassed by their love of data collection and analysis. The merit of these judges is excellently defended in 3.5.2

噸他炸薯?xiàng)l是由能夠RETINA.ai隊(duì)，他們的愛(ài)薯?xiàng)l只能由他們的數(shù)據(jù)收集和分析的熱愛(ài)超過(guò)了判斷。 這些法官的功績(jī)?cè)?.5.2中得到了很好的辯護(hù)

2.結(jié)果 (2. Results)

After the group play stage, and each contestant had been graded on the F.A.T System, we produce the following data table below:

一壓腳提升的小組賽階段，每名參賽者已分級(jí)的FAT系統(tǒng)，我們生產(chǎn)的以下數(shù)據(jù)見(jiàn)下表：

Those highlighted in green received acceptably high enough F.A.T scores to qualify for the tournament round. Those not highlighted fell below the McDonald’s line, named for the fact that not only was McDonald’s the lowest scored fry that qualified, but also with an average score of 2.43/5 McDonald’s can be considered an average, mediocre fry.

以綠色突出顯示的那些獲得了足夠高的FAT分?jǐn)?shù)，有資格參加比賽。 那些未突出顯示的部分不屬于麥當(dāng)勞產(chǎn)品線 ，因?yàn)辂湲?dāng)勞不僅是合格的得分最低的魚(yú)苗，而且平均得分為2.43 / 5麥當(dāng)勞也可以被認(rèn)為是平均的，平庸的魚(yú)苗。

It should also be noted that Little Ruby had an absolutely incredible score, scoring nearly a half point higher than any other contender and a near perfect score of 4.44/5. An additional interesting result was that Little Ruby and Burger Lounge scored higher than Great American Fries, a food truck that exclusively serves French fries.

還應(yīng)注意，Little Ruby的得分絕對(duì)令人難以置信，比其他競(jìng)爭(zhēng)者得分高近半點(diǎn)，接近完美得分為4.44 / 5。 另一個(gè)有趣的結(jié)果是，Little Ruby和Burger Lounge的得分高于僅供應(yīng)法式炸薯?xiàng)l的食品卡車(chē)Great American Fries。

All contestants above the McDonald’s line were entered the seeded tournament play resulting in the tournament structure seen in Figure 3. False identities are used because this is the same bracket judges saw during the contest.

麥當(dāng)勞線以上的所有參賽者都進(jìn)入了帶種子的錦標(biāo)賽比賽，結(jié)果如圖3所示。使用了錯(cuò)誤的身份，因?yàn)檫@與裁判在比賽中所看到的相同。

Figure 3: Qualifying Fries seeded for Tournament Play Bracket圖3：排位賽比賽排位賽的合格薯?xiàng)l

After the first round, Los Pollos Hermanos & Mos Eisley won as projected, with Los Pollos Hermanos sweeping JJ’s Diner 7–0. However, in the bottom two matchups the both fry-offs were upsets. Krusty Krab won decisively with a 6–1, while Bob’s Burgers squeaked out a victory 4–3 in a hotly debated heat. Krusty Krab knocking out Leaky Cauldron, and doing so effectively, is very surprising because as stated in the previous section, Leaky Cauldron (Great American Fries) only sells fries — and yet they couldn’t win a single game in the playoffs.

在第一輪之后， Los Pollos Hermanos和Mos Eisley贏得了預(yù)期， Los Pollos Hermanos以7-0橫掃了JJ的晚餐 。 但是，在最后兩場(chǎng)比賽中，兩次比賽都讓人沮喪。 Krusty Krab以6-1贏得了決定性的勝利，而B(niǎo)ob的Burgers在激烈爭(zhēng)論的熱火中以4-3擊敗了勝利。 Krusty Krab淘汰Leaky Cauldron并有效地做到這一點(diǎn)非常令人驚訝，因?yàn)槿缜耙还?jié)所述， Leaky Cauldron (美國(guó)大薯?xiàng)l) 只出售薯?xiàng)l-但他們無(wú)法在季后賽中贏得一場(chǎng)比賽。

The Final Four went as expected, with #1 Los Pollo Hermanos beating #4 Mos Eisley Cantina and #6 Krusty Krab dominating #7 Bob’s Burger 7–0. The Championship game was an instant ESPN classic. Seed #6 Krusty Krab, with only 1 vote against it the entire tournament, took on Seed #1 Los Pollos Hermanos, whose group play score was significantly higher than anyone else — and 1.44 points above their championship opponent. But in the end, after intense debate and judging, the underdog took the championship 4–3.

最終四強(qiáng)如期而至 ，第一名的Los Pollo Hermanos擊敗了第四名的Mos Eisley Cantina和第六名的Krusty Krab擊敗了第七名的 Bob's Burger 7-0。 冠軍賽是ESPN的經(jīng)典瞬間。 種子＃6 Krusty Krab在整個(gè)比賽中只有1票反對(duì)。他獲得了種子＃1 洛斯·波洛斯·赫曼諾斯(Los Pollos Hermanos )的團(tuán)體比賽成績(jī)，比其他任何人都高-并比冠軍對(duì)手高1.44分。 但是最后，經(jīng)過(guò)激烈的辯論和判斷，失敗者獲得了4-3的冠軍。

Hi Ho Burger (Krusty Krab) was crowned victor of Starch Madness!

Hi Ho Burger( Krusty Krab )被加冕為Starch Madness的勝利者！

In an amazing underdog story, Krusty Krab (HiHo Burger) wins Starch Madness despite being 6th seed!在令人驚嘆的失敗者故事中，Krusty Krab(HiHo Burger)盡管獲得了第六名，但還是贏得了Starch Madness！

3.討論 (3. Discussion)

Now that we have objective scores for each fry, we are given the opportunity to see the impact of certain attributes on French fry quality, derived from F.A.T score averages

?流，我們有客觀的分?jǐn)?shù)為每個(gè)魚(yú)苗，我們有機(jī)會(huì)看到某些屬性的法國(guó)魚(yú)苗質(zhì)量的影響，從FAT得分平均值得出

3.1 Does Price Imply Quality?

3.1價(jià)格是否暗示質(zhì)量？

Figure 4: The Relationship between Price and Fry Quality圖4：價(jià)格和油炸品質(zhì)之間的關(guān)系

When looking at French fry price vs quality, there is one thing that sticks out quite clearly: the most expensive French fry in the contest, The Misfit, was also voted the worst. In fact, the Misfit is such an outlier that inclusion of it as a data point actually changes the relationship between fry price and quality from positive to negative.

在查看炸薯?xiàng)l價(jià)格與質(zhì)量的對(duì)比時(shí)，有一件事情很明顯：在比賽中最昂貴的炸薯?xiàng)lThe Misfit也被評(píng)為最差。 實(shí)際上，Mifit異常嚴(yán)重，以至于將其作為數(shù)據(jù)點(diǎn)實(shí)際上改變了魚(yú)苗價(jià)格與質(zhì)量之間的關(guān)系，從正面變?yōu)樨?fù)面。

The correlation between price and quality without the Misfit is moderately positive (0.40 Pearson Correlation Coefficient) and the correlation with the Misfit is none to negative (-0.08 Pearson). To sum this up nicely:

沒(méi)有錯(cuò)配的價(jià)格與質(zhì)量之間的相關(guān)性為中等正值(0.40皮爾遜相關(guān)系數(shù))，而與錯(cuò)配的相關(guān)性為零至負(fù)(-0.08皮爾遜)。 總結(jié)一下：

• Price does imply quality, but only up to $10
  價(jià)格確實(shí)暗示著質(zhì)量，但最高不超過(guò)10美元
• The Misfit (and perhaps any/all fries that are above $10) are a rip off. You are better off going to McDonald’s
  Misfit(可能還有任何/所有高于10美元的炸薯?xiàng)l)都是騙子。 你最好去麥當(dāng)勞

Speaking of McDonalds — the cheapest option is actually the best option for the price you get. While absolute quality does increase as price increases, it does not increase to the point that you increase the quality per dollar spent. In fact, a dollar spent at McDonald’s produces 38% more fry quality than the next restaurant. If you’re looking for the best “bang for your buck,” McDonald’s is the best option.

說(shuō)到麥當(dāng)勞-最便宜的選擇實(shí)際上是您所獲得價(jià)格的最佳選擇。 雖然絕對(duì)質(zhì)量會(huì)隨著價(jià)格的上漲而增加，但不會(huì)增加到您提高每美元花費(fèi)的質(zhì)量。 實(shí)際上，在麥當(dāng)勞花一美元，比下一家餐館生產(chǎn)的魚(yú)苗質(zhì)量高38％。 如果您正在尋找最好的“物有所值”產(chǎn)品，那么麥當(dāng)勞是最佳選擇。

3.2 Does Distance imply Quality?

3.2距離是否暗示質(zhì)量？

A major discussion amongst judges during scoring was whether the distance from our office implied if the fries would be better or worse. The reasoning behind this is simply because fries are best when they are fresh and hot, and if we had to walk them back to the office for several minutes, they could lose more heat than a batch procured closer to our office.

評(píng)分過(guò)程中，評(píng)委們之間的主要討論是，與薯?xiàng)l之間的距離是否暗示了薯?xiàng)l的好壞。 這背后的原因僅是因?yàn)檎ㄊ項(xiàng)l在新鮮和熱的時(shí)候是最好的，如果我們不得不將它們帶回辦公室?guī)追昼?#xff0c;它們可能比在我們辦公室附近采購(gòu)的一批散發(fā)更多的熱量。

There are strong arguments for both sides. In support of this theory, Umami burger is one of the farthest and lowest rated fries, Little Ruby, Plan Check, and HiHo all made the playoffs and are very close to our office. However, Misfit, the closest restaurant, is the lowest score, and Great American Fries is far but still very high quality. Clearly, the only way to solve this is with

雙方都有很強(qiáng)的論點(diǎn)。 支持這一理論的是，鮮味漢堡是薯?xiàng)l中分離度最高和評(píng)分最低的一種，Little Ruby，Plan Check和HiHo都進(jìn)入了季后賽，并且非?？拷覀兊霓k公室。 但是，最近的餐廳Misfit的得分最低，而Great American Fries的質(zhì)量雖然很高，但仍然很高。 顯然，解決此問(wèn)題的唯一方法是

Statistics!

統(tǒng)計(jì)！

Location proximity and its Relationship to Fry Quality. Quadrants named moving clockwise from top-left位置鄰近性及其與油炸品質(zhì)的關(guān)系。 象限命名為從左上方順時(shí)針移動(dòng)

We charted where each restaurant was and found the Euclidean distance from each to our office, marked on the map. Firstly, even by removing The Misfit from the data, only 10% of variance can be explained through a linear or exponential relationship (0.1 R2 value), so the hypothesis that distance degrades quality is rejected. What the distance-quality chart does produce is an interesting bucketing strategy for these restaurants. By bisecting the data by both the McDonald’s line and the average distance, we can create 4 quadrants/segments.

我們繪制了每個(gè)餐廳的位置圖，并在地圖上找到了每個(gè)餐廳到我們辦公室的歐幾里得距離。 首先，即使通過(guò)從數(shù)據(jù)中刪除“失配”，也只能通過(guò)線性或指數(shù)關(guān)系(0.1 R2值)來(lái)解釋10％的方差，因此拒絕了距離降低質(zhì)量的假設(shè)。 距離質(zhì)量圖的結(jié)果確實(shí)是這些餐館的有趣存儲(chǔ)策略。 通過(guò)將數(shù)據(jù)按麥當(dāng)勞線和平均距離平分，我們可以創(chuàng)建4個(gè)象限/段。

Quadrant I can be labelled “Easy Bets,” where each restaurant is close and of high quality.

象限I可以標(biāo)記為“輕松下注”，其中每家餐廳都很近且質(zhì)量很高。

Quadrant II can be labelled “Worth the Journey” since they are farther away, but still high quality.

象限II可以標(biāo)記為“世界之旅”，因?yàn)樗鼈兙嚯x較遠(yuǎn)，但質(zhì)量仍然很高。

Quadrant III can be labelled “Not Worth the Journey” for having below-McDonalds’ quality at above average distance.

象限III因在高于平均距離的情況下具有低于麥當(dāng)勞的質(zhì)量而被標(biāo)記為“不值得一游”。

Finally Quadrant IV can be labelled “Deceptively Bad” because while they are close to the office, they are not even the short trip. The Quadrant IV name also stems from the fact that we frequent the Misfit often since it is so close, and we have grown to believe that their fries are quite good. Our studies show this is incorrect, and our frequent visits have simply biased us

最終， 象限IV可以標(biāo)記為“看似不好”，因?yàn)楸M管它們靠近辦公室，但路程也不短。 象限IV的名稱還源于這樣一個(gè)事實(shí)，即我們經(jīng)常會(huì)因?yàn)镸isfit如此之近而經(jīng)常去它，并且我們逐漸相信它們的薯?xiàng)l非常好。 我們的研究表明這是不正確的，而我們的頻繁拜訪只是使我們有偏見(jiàn)

3.3 Did External Factors influence our Voting?

3.3外部因素是否影響了我們的投票？

3.3.1 Comedic False Identity

3.3.1喜劇虛假身份

The false identities were added to the restaurants so that people wouldn’t be biased based on their prior experiences with that restaurant when voting. However, in our attempt to mask their identities, we may have introduced new bias through our associations with the cinematic universes that the fictional restaurants are from. Specifically, what if a restaurant was given a comedic identity, and because of our positive association with that movie/show, we graded it higher? Or conversely, what if a restaurant was given a very serious movie/show identity, and because of this, the judges scrutinized the fry more closely? Would people treat associations to Silence of the Lambs more negatively than Dumb and Dumber?

將虛假身份添加到餐廳，以便人們?cè)谕镀睍r(shí)不會(huì)因以前在該餐廳的經(jīng)歷而產(chǎn)生偏見(jiàn)。 但是，為了掩蓋他們的身份，我們可能通過(guò)與虛構(gòu)餐廳所來(lái)自的電影世界的聯(lián)系引入了新的偏見(jiàn)。 具體來(lái)說(shuō)，如果一家餐廳被賦予喜劇身份，又由于我們與那部電影/節(jié)目有著積極的聯(lián)系，我們會(huì)給它更高的評(píng)分嗎？ 或相反，如果給餐廳以非常嚴(yán)肅的電影/表演身份，又因?yàn)檫@個(gè)原因，法官會(huì)更仔細(xì)地檢查炸薯?xiàng)l怎么辦？ 人們會(huì)不會(huì)比沉默寡言和沉默寡言對(duì)待與沉默的羔羊相關(guān)的協(xié)會(huì)？

The answer is no. Splitting fries by comedy, we get a t-test p value of 0.52. The comedic value of the false identity did not influence our voting.

答案是否定的 。 按喜劇拆分炸薯?xiàng)l，我們得到的t檢驗(yàn)p值為0.52。 虛假身份的喜劇價(jià)值不影響我們的投票。

3.3.2 Weather

3.3.2天氣

We conducted these experiments during February and March — the heart of winter in Los Angles. It reached a chilling 58 degrees during some of our rounds — did this influence our votes? Did the cold weather make the warm fries taste even better, and therefore result in higher F.A.T Scores? Or did the overcast weather make us somber, and therefore more critical of the fries we ate?

我們?cè)?月和3月(洛杉磯的冬季中心)進(jìn)行了這些實(shí)驗(yàn)。 在我們的某些回合中，溫度達(dá)到了令人震驚的58度-這是否影響了我們的選票？ 寒冷的天氣是否使熱炸薯?xiàng)l的味道更好，從而導(dǎo)致FAT分?jǐn)?shù)更高？ 還是因?yàn)殛幱晏鞖馐刮覀儜n郁，因此對(duì)我們吃的薯?xiàng)l更加挑剔？

The answer is no. Splitting fries by if the weather as above 60 degrees of not, we get a t-test p value of 0.74. The weather outside did not influence our voting

答案是否定的 。 如果天氣不高于60度，則將薯?xiàng)l分開(kāi)，我們得到的t檢驗(yàn)p值為0.74。 外面的天氣沒(méi)有影響我們的投票

3.3.3 Day of the Week

3.3.3星期幾

Does judging fries on Fryday impact our votes? No, a p value of 0.55 rejects this hypothesis.

在Fryday進(jìn)行炸薯?xiàng)l評(píng)判是否會(huì)影響我們的投票？ 不 ，p值為0.55否定了這一假設(shè)。

3.4 Can Yelp Ratings indicate F.A.T Score?

3.4 Yelp等級(jí)可以表示FAT分?jǐn)?shù)嗎？

Yelp is a service that prides itself on crowd-sourced ratings for restaurants. Many often check Yelp first before deciding if they should eat a particular location. Can we use the same strategy when selecting a location to dine on the best French fries? One would hope that as the overall Yelp score increases, so does the quality of the French fry.

Y elp是一家以餐廳的眾包評(píng)分為傲的服務(wù)。 許多人通常會(huì)先檢查Yelp，然后再?zèng)Q定是否應(yīng)該在特定地點(diǎn)用餐。 選擇最佳薯?xiàng)l用餐地點(diǎn)時(shí)，可以使用相同的策略嗎？ 人們希望隨著Yelp總體得分的提高，炸薯?xiàng)l的質(zhì)量也提高。

Figure 8: The Relationship between Yelp Scores and Fry Quality圖8：Yelp得分與油炸品質(zhì)之間的關(guān)系

Unfortunately, this is not the case. As seen in Figure 8, highly rated Yelp eaters fail to produce on average fries that are as highly rated as their restaurant. Even when we cut out the Misfit outlier, good Yelp restaurants fall below the Yelp-FAT Line, which simply charts a 1:1 relationship between Yelp score and F.A.T score — a restaurant above the Yelp-FAT line (such as 3.0 Yelp score) produces fries of a higher caliper that the Yelp score would suggest. Our explanation for the 4.0+ restaurants failing to produce 4.0+ fries is that the Yelp score includes not only other foods, but also the ambiance of a restaurant. High Yelp score restaurants tend to have a better dining experience than lower scoring locations, despite their fries being of an equal or lesser quality.

不幸的是，這種情況并非如此。 如圖8所示，評(píng)分較高的Yelp食用者平均生產(chǎn)的薯?xiàng)l不如其餐廳。 即使我們剔除了Misfit異常值，良好的Yelp餐館也會(huì)落在Yelp-FAT線以下，該圖表簡(jiǎn)單地繪制了Yelp分?jǐn)?shù)與FAT分?jǐn)?shù)之間的1：1關(guān)系-一家餐館在Yelp-FAT線之上(例如3.0 Yelp分?jǐn)?shù))產(chǎn)生了更高的卡尺炸薯?xiàng)l，這是Yelp得分所暗示的。 我們對(duì)4.0+餐館未能生產(chǎn)4.0+薯?xiàng)l的解釋是，Yelp評(píng)分不僅包括其他食物，還包括餐館的氛圍。 盡管炸薯?xiàng)l的質(zhì)量相同或較低，但高Yelp評(píng)分的餐廳往往比低分的餐廳擁有更好的就餐體驗(yàn)。

3.5 Addressing Potential Confounding Variables

3.5解決潛在的混雜變量

3.5.1 Playoff Rankings

3.5.1季后賽排名

Everyone loves an underdog story, and so it is possible that despite having blinded each restaurant during group and tournament play, knowing the rankings could have compelled voters to choose the underdog when a contest was close. Fry quality in tournament play was higher than in group play, and so this is entirely possible. How to control for this in the future is to avoid telling rankings, but also changing the false identities several times throughout the contest, thereby removing any bias judges have procured during previous tastings.

?veryone愛(ài)一個(gè)失敗者的故事，所以有可能的是，盡管已經(jīng)蒙蔽組錦標(biāo)賽在每個(gè)餐廳，知道的排名可能迫使選民選擇弱旅當(dāng)比賽接近。 錦標(biāo)賽比賽中的魚(yú)苗質(zhì)量要高于團(tuán)體比賽中的魚(yú)苗質(zhì)量，因此這完全有可能。 未來(lái)如何控制這一點(diǎn)是為了避免透露排名，而且還要在整個(gè)比賽中多次更改錯(cuò)誤的身份，從而消除法官在先前的品嘗中所造成的偏見(jiàn)。

3.5.2 Judge Acumen

3.5.2明智的法官

To address the concern that the judges themselves were bias and not indicative of scores of those not involved in the research, this is also possible. However, each judge has passed the French Fry Judge Certification Program (Figure 9) and therefore their expertise should be considered adequately accurate for French fry quality.

為了解決法官本身有偏見(jiàn)，并不能表明未參與研究的人員的分?jǐn)?shù)的擔(dān)憂，這也是可能的。 但是，每個(gè)法官都通過(guò)了法國(guó)炸薯?xiàng)l法官認(rèn)證計(jì)劃(圖9)，因此，他們的專業(yè)知識(shí)對(duì)于法國(guó)炸薯?xiàng)l質(zhì)量應(yīng)被認(rèn)為足夠準(zhǔn)確。

Figure 9: Logo of Certification Program all Judges received圖9：所有收到評(píng)審的認(rèn)證計(jì)劃徽標(biāo)

4。結(jié)論 (4. Conclusion)

Ranking French Fries is a relevant subject to all readers, and we hope to continue our research beyond the Santa Monica area. While our resources confined us to Santa Monica and single elimination, next steps could be an expansion of both region and trials. We are all interested if these results hold up across all restaurant locations, or if there are locations that claim to have the best French Fry that wish to enter the challenge.

[R安慶炸薯?xiàng)l是相關(guān)學(xué)科的所有讀者，我們希望繼續(xù)我們的研究超越了圣莫尼卡地區(qū)。 盡管我們的資源將我們限制在圣莫尼卡和一次淘汰賽，但下一步可能是擴(kuò)大地區(qū)和擴(kuò)大試驗(yàn)范圍。 我們都感興趣的是，這些結(jié)果是否在所有餐廳位置都得到了證明，或者是否有聲稱擁有最好炸薯?xiàng)l的位置希望挑戰(zhàn)。

For those unconvinced that the research topic of Fry Quality of Santa Monica French Fries was worthy of publication, we would strongly argue that the pursuit of knowledge should always be encouraged, no matter how small the impact or specialized the application.

對(duì)于那些不相信《圣莫尼卡炸薯?xiàng)l品質(zhì)》的研究主題值得發(fā)表的人們，我們強(qiáng)烈認(rèn)為，無(wú)論影響有多小或應(yīng)用多么專業(yè)，都應(yīng)始終鼓勵(lì)對(duì)知識(shí)的追求。

We hope that all those who read it feel empowered to conduct their own French fry quality assessment as a form of peer review, and publish any interesting results found. The research area of French Fry Quality is young and ripe for more publications.

我們希望所有閱讀它的人 感到有能力以同行評(píng)審的形式進(jìn)行自己的薯?xiàng)l質(zhì)量評(píng)估，并發(fā)布發(fā)現(xiàn)的有趣結(jié)果。 炸薯?xiàng)l質(zhì)量研究領(lǐng)域很年輕，可以發(fā)表更多的出版物。

The Quest for the Ultimate French Fry is still afoot, and Data Scientists have a major role to play in the road ahead.

最終炸薯?xiàng)l的探索仍在進(jìn)行中，數(shù)據(jù)科學(xué)家在未來(lái)的道路上將發(fā)揮重要作用。

Matthew Kalapuch on Matthew Kalapuch在未Unsplash飛濺

Hope you enjoyed,

希望你喜歡，

Adam “Fry” Brownell

亞當(dāng)·弗萊·布朗內(nèi)爾

翻譯自: https://towardsdatascience.com/french-fries-the-final-frontier-of-data-science-674ec8f1d74c

kfc流程管理炸薯?xiàng)l幾秒

總結(jié)

以上是生活随笔為你收集整理的kfc流程管理炸薯条几秒_炸薯条成为数据科学的最后前沿的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問(wèn)題。

如果覺(jué)得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。