Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/thammami01/simple-recruitment-ml

Simple recruitment app that allows job posting/application, and viewing regression/classification figures based on entries.
https://github.com/thammami01/simple-recruitment-ml

flask matplot-lib mongodb python scikit-learn

Last synced: about 18 hours ago
JSON representation

Simple recruitment app that allows job posting/application, and viewing regression/classification figures based on entries.

Awesome Lists containing this project

README 

        
# Recruitment ML



## Introduction



This project is a simple web app dedicated to recruitment. It allows companies to post job offers, candidates to apply, and recruiters to view regression/classification figures based on salary suggestions, years of experience, and possession of a degree.



## Technologies



- Python: The main programming language for the backend and data analysis.

- Flask: A Python web framework used to develop the backend of the application.

- MongoDB: A NoSQL database used to store data related to job offers and applications.

- pymongo: A Python library for interacting with MongoDB from the Flask application.

- Marshmallow: A Python library used for schema validation and data serialization.

- scikit-learn: A Python library for machine learning tools, used to generate regression and classification figures.

- Matplotlib: A Python library for creating graphs, used to visualize the generated figures.

- NumPy: A Python library for numerical calculations, often used in conjunction with scikit-learn.



These technologies were chosen for their efficiency in software development, data manipulation, and creating machine learning models.



## Tables and Relationships



| Job | Application |

|-----|-----|

| id `PK/uid` | id `PK/uid` |

| title `string` | years_of_experience `number` |

| company `string` | has_diploma `boolean` |

| workspace_type `string` | salary `number` |

| location `string` | email `string` |

| type `string` | cv_url `string` |

| description `string` | job_id `FK/string` |

| email `string` |



`Job.id` 1 - 0..* `Application.job_id`



## UIs



The user interfaces are those of the job posting process on LinkedIn.



1. Job Details

01



2. Job Description

02



3. Pre-screening Questions

03



## Backend Endpoints



- `/jobs`:

  - GET `/`:

    Retrieves a list of available jobs from the database.

  - POST `/`:

    Creates a new job posting.

  - POST `//apply`:

    Allows candidates to submit their applications by providing details such as name, years of experience, possession of a degree, salary expectations, email, and a link to their CV.

  - GET `//stats`:

    Retrieves statistics and generates machine learning figures based on the received applications for a specific job. The figures include regression and classification visualizations based on suggested candidates entries (salaries, years of experience, and possession of a degree).



- `/figures`:

  - GET `/`:

    Retrieves the image file of a specific machine learning figure identified by its unique ID.



Example of response using an API client:



04



Example data within the Jobs collection:



05



Example data within the Applications collection:



06



## Figures Types



Three types of figures have been employed in the application:



- Decision Tree for Classification:

The first type of figure showcases a decision tree based on classification data. This tree provides a visual representation of classification criteria, such as years of experience, suggested salaries, and possession of a degree. The nodes of the tree illustrate the decisions made to classify applications based on these criteria.



- Decision Tree for Regression:

The second type presents a decision tree for regression. It is designed to assess the relationships between years of experience and suggested salaries.



- Linear Regression:

In addition to decision trees, we will generate linear regression plots. This third type of figure presents a linear regression line that models the relationship between a specific feature (such as years of experience) and the target variable (suggested salaries). This graphical representation provides an intuitive view of regression and facilitates understanding of linear relationships between variables.



## Figures Visualization



Sample job applications used for demonstration:



```python

[

  {"salary": 1000, "years_of_experience": 0, "has_diploma": False},

  {"salary": 1500, "years_of_experience": 1, "has_diploma": False},

  {"salary": 1700, "years_of_experience": 1, "has_diploma": True},

  {"salary": 1200, "years_of_experience": 0, "has_diploma": False},

  {"salary": 1150, "years_of_experience": 0, "has_diploma": False},

  {"salary": 1800, "years_of_experience": 3, "has_diploma": True},

  {"salary": 1550, "years_of_experience": 1, "has_diploma": False},

  {"salary": 1750, "years_of_experience": 2, "has_diploma": False},

  {"salary": 2200, "years_of_experience": 4, "has_diploma": True},

  {"salary": 2050, "years_of_experience": 5, "has_diploma": True},

  {"salary": 3000, "years_of_experience": 5, "has_diploma": True},

  {"salary": 4000, "years_of_experience": 10, "has_diploma": True},

  {"salary": 3500, "years_of_experience": 8, "has_diploma": True},

  {"salary": 4500, "years_of_experience": 10, "has_diploma": True},

  {"salary": 4200, "years_of_experience": 12, "has_diploma": True},

]

```



We will integrate these three types of figures on eight occasions:



- Decision Tree - Experience (feature) vs Salary (target):

A decision tree based on candidates' years of experience, providing insights into how experience influences suggested salaries.



07



- Decision Tree - Salary vs Experience:

A second decision tree based on suggested salaries, showing how salaries are influenced by candidates' experience.



08



- Decision Tree - Degree vs Experience:

A decision tree analyzing the relationship between possessing a degree and candidates' years of experience.



09



- Decision Tree - Degree vs Salary:

Another decision tree exploring how possessing a degree influences suggested salaries.



10



- Decision Tree - Experience vs Degree:

A decision tree classifying applications based on experience and indicating the relationship with possessing a degree.



11



Explanation: the value = [3, 0] indicates that there are 0 samples with the target class (has a degree) True and 3 samples with the target class False. This means that 3 do not have degrees in the field, and 0 do.



- Decision Tree - Salary vs Degree:

A decision tree analyzing the relationship between suggested salaries and possessing a degree.



12



- Linear Regression - Experience vs Salary:

A linear regression plot illustrating the relationship between experience and suggested salaries, with a prediction range from 0 to 20 years of experience.



13



- Linear Regression - Salary vs Experience:

Another linear regression plot presenting the inverse relationship, this time between suggested salaries and candidates' experience, with a salary range from 1000 to 5000.



14