Women In AI Hackathon: MachineHack Presents ‘Food Quality Assessment’

MachineHack is back with an exciting hackathon, and this time, we challenge the women in the analytics space. As Analytics India Magazine prepares itself for the 2nd Edition of the Women in Analytics Conference – The Rising 2020, we are looking forward to the women data scientists to participate in this exciting hackathon to stand a chance to win free passes to the conference. ‘The Rising’ is a platform where leading women visionaries will dive into the buzzing field of data science and machine learning and share their perspective on how to build a perfect career.

Problem Description

The food inspection department conducts regular inspections of food quality for various restaurants in the city. It’s a very well documented procedure, and over time some good amount of data has been generated out of these inspections. 

The inspection department would like to predict where they should focus most in terms of their next inspection schedule so that they can most optimize their time at hand to catch the worst offenders. Can the past inspection or any data that they have collected predict which facility will pass or fail.

---

---

In this hackathon, MachineHack will provide you with a subset of this dataset with information on food quality checks conducted on thousands of facilities that serve food across multiple cities. Your objective as a data scientist is to predict whether a facility will pass or fail the inspection based on several factors.

Objective

Build a predictive model that is capable of predicting the outcome of an inspection conducted in a facility based on the given set of features.

Features : 

ID: A unique ID for each inspection

Date: The date at which the inspection was done in a particular facility

License No: De-identified license number for a particular facility

Facility ID: De-identified unique facility id  for a facility

Facility Name: The encoded name of a facility

Type: The type of the facility being inspected

Street: The encoded street where the facility is located

City: The encoded city where the facility is located

State: The encoded state where the facility is located

Location ID: An encoded location feature.

Reason: The primary reason for the inspection

Section Violations: Laws violated by the facility

Risk Level: The level of risk the facility possesses to the consumers.

Geo_Loc: De-identified geolocation of the facility

Inspection_Results: The result of the inspection

Target Values:

The actual inspection results and their encoded variables are given below: 

• 0: ’FACILITY CHANGED’
• 1: ’FAIL’
• 2: ’FURTHER INSPECTION REQUIRED’,
• 3: ’INSPECTION OVERRULED’
• 4: ’PASS’
• 5: ’PASS(CONDITIONAL)’
• 6: ’SHUT-DOWN’

Here Is What The Data Looks Like:

Bounties

Top 3 women competitors will win free passes to ‘The Rising  2020’.

To know more about The Rising 2020, click here.

Rules

1. Top three women competitors will be eligible for bounties
2. This is an individual competition, and team participation is not allowed 
3. One account per participant. Submissions from multiple accounts will lead to disqualification.
4. Users may submit the solution any number of times before the specified hackathon expiry date.
5. You will not be able to submit once you click the “Complete Hackathon” button. Do not ignore this feature.
6. This hackathon is open to all registered users who are 18 years or older
7. This competition counts towards our overall ranking points.
8. We ask that you respect the spirit of the competition and do not cheat.

Timeline

January 31, 2020 to March 10, 2020

The winners will be finalized on March 10, 2020, at 6 PM IST based on the leaderboard standings.

Evaluation

The leaderboard is evaluated using Multi-Class Log loss (Cross-entropy loss) for the participant’s submission.