New York Pizza How to Find the Best Jared P. Lander Columbia - - PowerPoint PPT Presentation

▶
new york pizza
SMART_READER_LITE
LIVE PREVIEW

New York Pizza How to Find the Best Jared P. Lander Columbia - - PowerPoint PPT Presentation

Mar 22, 2023 110 likes •469 views

New York Pizza How to Find the Best Jared P. Lander Columbia University December 3, 2008 What is New York Style? What is New York Style? Charring and air bubbles around the crust What is New York Style? Thin Slices Charring and air

slide-1
SLIDE 1

New York Pizza

How to Find the Best Jared P. Lander Columbia University December 3, 2008

slide-2
SLIDE 2

What is New York Style?

slide-3
SLIDE 3

What is New York Style?

Charring and air bubbles around the crust

slide-4
SLIDE 4

What is New York Style?

Charring and air bubbles around the crust Thin Slices

slide-5
SLIDE 5

What is New York Style?

Charring and air bubbles around the crust Thin Slices Just the right bend for folding

slide-6
SLIDE 6

Family Tree

slide-7
SLIDE 7

Family Tree

Lombardi’s (1905)

slide-8
SLIDE 8

Family Tree

Lombardi’s (1905) Totonno’s (1924)

slide-9
SLIDE 9

Family Tree

Lombardi’s (1905) Totonno’s (1924) John’s of Bleecker (1929)

slide-10
SLIDE 10

Family Tree

Lombardi’s (1905) Totonno’s (1924) Patsy’s (1933) John’s of Bleecker (1929)

slide-11
SLIDE 11

Family Tree

Lombardi’s (1905) Totonno’s (1924) Patsy’s (1933) John’s of Bleecker (1929) Grimaldi’s (1990)

slide-12
SLIDE 12

The Data

  • MenuPages.com has menus, ratings, locations and price information for

nearly 700 restaurants tagged as pizza in Manhattan and parts of Brooklyn

  • Slice, a pizza blog, has a definitive listing (and maps) of coal pizzerias
  • An exhaustive Google search provided wood classifications
slide-13
SLIDE 13

The Data

slide-14
SLIDE 14

Rating Histogram

slide-15
SLIDE 15

Ratings Breakdown

slide-16
SLIDE 16

Model for Rating

  • logit(Rating/5.5) ~ 0.77571 – 0.1782*AreaMidtown
  • Coefficients:
  • Estimate SE t-value Pr(>|t|)
  • (Intercept) 0.60326 0.20090 3.003 0.00278 **
  • AreaMidtown -0.17801 0.06628 -2.686 0.00743 **
slide-17
SLIDE 17

Model for Rating

  • logit(Rating/5.5) ~ 0.77571 – 0.1782*AreaMidtown
  • Coefficients:
  • Estimate SE t-value Pr(>|t|)
  • (Intercept) 0.60326 0.20090 3.003 0.00278 **
  • AreaMidtown -0.17801 0.06628 -2.686 0.00743 **
  • AreaDowntown-0.05766 0.05527 -1.043 0.29724
  • AreaUptown -0.04838 0.06016 -0.804 0.42154
  • FuelGas 0.13354 0.11575 1.154 0.24905
  • FuelWood 0.07606 0.12980 0.586 0.55810
  • PriceExpensive-0.02827 0.04009 -0.705 0.48097
  • ReviewsLow 0.03452 0.17958 0.192 0.84762
  • ReviewsMedium0.02728 0.19083 0.143 0.88638
  • PizzaName 0.02684 0.06227 0.431 0.66661
slide-18
SLIDE 18

Model for Rating

  • logit(Rating/5.5) ~ 0.77571 – 0.1782*AreaMidtown
  • Wald test
  • Model 1: logit(Rating/5.5) ~ Area
  • Model 2: logit(Rating/5.5) ~ 1
  • Res.Df

Df F Pr(>F)

  • 1 640
  • 2 643
  • 3

2.8195 0.03826 *

  • Perhaps is more appropriate due to its simplicity and the

minimal impact of the lone coefficient in the model

  • Maybe all pizza is for the most part pretty good
slide-19
SLIDE 19

Ratings Breakdown

.020 .005 .006 One Sided P-Value

slide-20
SLIDE 20

Review Count

Use the number of reviews as a proxy for popularity Why do people frequent these pizzerias in the first place?

slide-21
SLIDE 21

Number of Reviews by Area

slide-22
SLIDE 22

Number of Reviews by Fuel

slide-23
SLIDE 23

Number of Reviews by Price

slide-24
SLIDE 24

Reviews Histogram

Approximately Poisson

slide-25
SLIDE 25

Model for Review Count

  • Number.Reviews ~ 3.826 + 0.157*AreaUptown –

1.641*FuelGas – 0.734*FuelWood - 0.545*PriceExpensive + 0.119*FuelGas:PriceExpensive + 0.670*FuelWood:PriceExpensive

  • (Poisson Regression)
  • Coefficients:
  • Estimate SE z-value Pr(>|z|)
  • (Intercept)

3.82693 0.05899 64.876 < 2e-16 ***

  • AreaUptown 0.15733 0.02156 7.297 2.94e-13 ***
  • FuelGas -1.64163 0.06019 -27.273 < 2e-16 ***
  • FuelWood
  • 0.73476 0.08884 -8.271 < 2e-16 ***
  • PriceExpensive -0.54561 0.07919 -6.890 5.60e-12 ***
  • FuelGas:PriceExpensive 1.11908 0.08289 13.501 < 2e-16 ***
  • FuelWood:PriceExpensive 0.67017 0.10819 6.195 5.85e-10 ***
slide-26
SLIDE 26

Model for Review Count

  • Number.Reviews ~ 3.826 + 0.157*AreaUptown –

1.641*FuelGas – 0.734*FuelWood - 0.545*PriceExpensive + 0.119*FuelGas:PriceExpensive + 0.670*FuelWood:PriceExpensive

  • (Poisson Regression)
  • Wald test
  • Model 1: Number.Reviews ~ Area + Fuel + Price +

Price * Fuel

  • Model 2: Number.Reviews ~ 1
  • Res.Df Df F Pr(>F)
  • 1 637
  • 2 643 -6 309.94 < 2.2e-16 ***
slide-27
SLIDE 27

Fitted Values

Coal and then wood dominates! All pizzerias on this list are categorized as expensive Both Uptown and Downtown

slide-28
SLIDE 28

Most Popular

slide-29
SLIDE 29

Top Pizzerias

  • Lombardi’s
  • Di Fara
  • Joe’s
  • John’s of Bleecker
  • Totonno’s (Coney

Island)

  • Vinny Vincenz
  • Pizza 33
  • No. 28
  • Patsy’s (East

Harlem)

  • Artichoke
  • Maffei’s
  • New York Pizza

Suprema

  • Una Pizza

Napoletana

  • Franny’s
slide-30
SLIDE 30

Conclusions

slide-31
SLIDE 31

Conclusions

  • Coal ovens attract large

crowds and the higher prices are worth paying

slide-32
SLIDE 32

Conclusions

  • Coal ovens attract large

crowds and the higher prices are worth paying

  • The old adage about

pizza is true: “Even when it’s bad, it’s still good.”

slide-33
SLIDE 33

Blogosphere

Slice Serious Eats Midtown Lunch Revolution Computing NBC New York

slide-34
SLIDE 34

Jared’s Picks

Pie: Margherita: Neapolitan: Upscale: Largest: John’s of Bleecker Pizza 33 Joe’s (Carmine St.) Kesté Koronet Specialty: Specialty: Grandma: Upside Down: Cheapest: Artichoke Vinny Vincenz Maffei’s New York Pizza Suprema Crocodile Lounge*

* Free with purchase of beer