Why Inspection Prediction? Public, Private, Commercial Good

As a gourmet, one of the largest concerns is the food security. From a macro perspective, both different facility owners and inspectors would love to know the result of their inspection so they can prepare appropriate resources for additional work or improvement.

1. Demand:

• People want to have decent, clean food
• Facilities’ owners
• inspectors at Chicago City

2. Function:

• Show location
• Dashboard control for RiskType, FacilitiesType
• Predict Outcome of Inspection

Data in Use

The data has been scrapped directly from the records of Chicago City Hall website. The original training data contains 199943 records in total. After the data manipulation, I restricted the conditions to four major elements: 1. Facility Type: School, Bakery, Restaurant, and school. 2. Risk TYpe: Risk 1(High), Risk 2(Medium), Risk 3(Low) 3. Location(Longtitude & Latitude) The summary is as follow:

  summary(main_data)

##  Results                  Risk         Latitude       Longitude     
##  fail:1178   risk 1 (high)  :4713   Min.   :41.65   Min.   :-87.91  
##  pass:4412   risk 2 (medium): 838   1st Qu.:41.85   1st Qu.:-87.71  
##              risk 3 (low)   :  39   Median :41.89   Median :-87.67  
##                                     Mean   :41.89   Mean   :-87.68  
##                                     3rd Qu.:41.94   3rd Qu.:-87.63  
##                                     Max.   :42.02   Max.   :-87.53  
##         Ftype     
##  bakery    :  96  
##  catering  :  40  
##  restaurant:4691  
##  school    : 763  
##                   
## 

“Chicago”

Empower Prediction by Logistic Regression

The basic model used here is the logstic model, regarding the existing independent variables, categorical and numeric. The model summary is as follow:

## 
## Call:
## glm(formula = Results ~ Ftype + Risk + Latitude + Longitude, 
##     family = binomial, data = main_data)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.8951   0.6164   0.6942   0.7014   0.8613  
## 
## Coefficients:
##                      Estimate Std. Error z value Pr(>|z|)   
## (Intercept)         -20.85360   46.09521  -0.452  0.65098   
## Ftypecatering         0.36929    0.46524   0.794  0.42734   
## Ftyperestaurant       0.27524    0.24528   1.122  0.26180   
## Ftypeschool           0.24201    0.26022   0.930  0.35236   
## Riskrisk 2 (medium)   0.26347    0.09881   2.667  0.00766 **
## Riskrisk 3 (low)     -0.20933    0.37100  -0.564  0.57260   
## Latitude             -0.24903    0.47078  -0.529  0.59682   
## Longitude            -0.36843    0.57176  -0.644  0.51933   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 5756.9  on 5589  degrees of freedom
## Residual deviance: 5747.6  on 5582  degrees of freedom
## AIC: 5763.6
## 
## Number of Fisher Scoring iterations: 4

The model performs effectively but requires time to run on the server at this stage.

Why this App?

Before showing the real app, hereby I provide a summary of the UI elements. The sidepanel includes 4 major inputs groups.

• Longtitude & Latitude: These inputs turn into the locator.
• Ftype: This inputs provides the facility type.
• Rtype: This inputs provides the risk type.
• Show Summary: This option enable users to hide the summary.

This app provides a convinent, powerful, reiliable predictor for any one interested in the subject.

Thank You & Lets try it.

Thank You and lets try it out.

• Links as follow: “https://ronngbinye94.shinyapps.io/RY_DPDW4_noicon/”

“Chicago Food”