Full-time. 1 minute read. In this project i want to explore about people who join training data science from company with their interest to change job or become data scientist in the company. Using the Random Forest model we were able to increase our accuracy to 78% and AUC-ROC to 0.785. Please AVP, Data Scientist, HR Analytics. There was a problem preparing your codespace, please try again. Many people signup for their training. Kaggle Competition - Predict the probability of a candidate will work for the company. I got -0.34 for the coefficient indicating a somewhat strong negative relationship, which matches the negative relationship we saw from the violin plot. I got my data for this project from kaggle. Answer In relation to the question asked initially, the 2 numerical features are not correlated which would be a good feature to use as a predictor. was obtained from Kaggle. The pipeline I built for prediction reflects these aspects of the dataset. In other words, if target=0 and target=1 were to have the same size, people enrolled in full time course would be more likely to be looking for a job change than not. Since our purpose is to determine whether a data scientist will change their job or not, we set the 'looking for job' variable as the label and the remaining data as training data. Github link all code found in this link. has features that are mostly categorical (Nominal, Ordinal, Binary), some with high cardinality. Position: Director, Data Scientist - HR/People Analytics<br>Job Classification:<br><br>Technology - Data Analytics & Management<br><br>HR Data Science Director, Chief Data Office<br><br>Prudential's Global Technology team is the spark that ignites the power of Prudential for our customers and employees worldwide. First, Id like take a look at how categorical features are correlated with the target variable. A tag already exists with the provided branch name. Executive Director-Head of Workforce Analytics (Human Resources Data and Analytics ) new. Underfitting vs. Overfitting (vs. Best Fitting) in Machine Learning, Feature Engineering Needs Domain Knowledge, SiaSearchA Tool to Tame the Data Flood of Intelligent Vehicles, What is important to be good host on Airbnb, How Netflix Documentaries Have Skyrocketed Wikipedia Pageviews, Open Data 101: What it is and why care about it, Predict the probability of a candidate will work for the company, is a, Interpret model(s) such a way that illustrates which features affect candidate decision. This dataset consists of rows of data science employees who either are searching for a job change (target=1), or not (target=0). Many people signup for their training. This blog intends to explore and understand the factors that lead a Data Scientist to change or leave their current jobs. If nothing happens, download Xcode and try again. Group Human Resources Divisional Office. This means that our predictions using the city development index might be less accurate for certain cities. Following models are built and evaluated. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Our mission is to bring the invaluable knowledge and experiences of experts from all over the world to the novice. Please So I performed Label Encoding to convert these features into a numeric form. Target isn't included in test but the test target values data file is in hands for related tasks. We used the RandomizedSearchCV function from the sklearn library to select the best parameters. Tags: The approach to clean up the data had 6 major steps: Besides renaming a few columns for better visualization, there were no more apparent issues with our data. Then I decided the have a quick look at histograms showing what numeric values are given and info about them. These are the 4 most important features of our model. as this is only an initial baseline model then i opted to simply remove the nulls which will provide decent volume of the imbalanced dataset 80% not looking, 20% looking. I do not own the dataset, which is available publicly on Kaggle. . StandardScaler is fitted and transformed on the training dataset and the same transformation is used on the validation dataset. Use Git or checkout with SVN using the web URL. Using the pd.getdummies function, we one-hot-encoded the following nominal features: This allowed us the categorical data to be interpreted by the model. to use Codespaces. Exploring the categorical features in the data using odds and WoE. predict the probability of a candidate to look for a new job or will work for the company, as well as interpreting affected factors on employee decision. 1 minute read. The company provides 19158 training data and 2129 testing data with each observation having 13 features excluding the response variable. which to me as a baseline looks alright :). MICE is used to fill in the missing values in those features. It is a great approach for the first step. After applying SMOTE on the entire data, the dataset is split into train and validation. This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Use Git or checkout with SVN using the web URL. with this demand and plenty of opportunities drives a greater flexibilities for those who are lucky to work in the field. If nothing happens, download Xcode and try again. Are you sure you want to create this branch? Hiring process could be time and resource consuming if company targets all candidates only based on their training participation. Please refer to the following task for more details: Recommendation: As data suggests that employees who are in the company for less than an year or 1 or 2 years are more likely to leave as compared to someone who is in the company for 4+ years. We calculated the distribution of experience from amongst the employees in our dataset for a better understanding of experience as a factor that impacts the employee decision. Each employee is described with various demographic features. Recommendation: The data suggests that employees with discipline major STEM are more likely to leave than other disciplines(Business, Humanities, Arts, Others). - Build, scale and deploy holistic data science products after successful prototyping. The dataset is imbalanced and most features are categorical (Nominal, Ordinal, Binary), some with high cardinality. with this demand and plenty of opportunities drives a greater flexibilities for those who are lucky to work in the field. All dataset come from personal information . Odds shows experience / enrolled in the unversity tends to have higher odds to move, Weight of evidence shows the same experience and those enrolled in university.;[. Apply on company website AVP, Data Scientist, HR Analytics . Thus, an interesting next step might be to try a more complex model to see if higher accuracy can be achieved, while hopefully keeping overfitting from occurring. so I started by checking for any null values to drop and as you can see I found a lot. Schedule. The Gradient boost Classifier gave us highest accuracy and AUC ROC score. For more on performance metrics check https://medium.com/nerd-for-tech/machine-learning-model-performance-metrics-84f94d39a92, _______________________________________________________________. Does more pieces of training will reduce attrition? For the full end-to-end ML notebook with the complete codebase, please visit my Google Colab notebook. Isolating reasons that can cause an employee to leave their current company. As we can see here, highly experienced candidates are looking to change their jobs the most. For this, Synthetic Minority Oversampling Technique (SMOTE) is used. In our case, the columns company_size and company_type have a more or less similar pattern of missing values. to use Codespaces. Github link: https://github.com/azizattia/HR-Analytics/blob/main/README.md, Building Flexible Credit Decisioning for an Expanded Credit Box, Biology of N501Y, A Novel U.K. Coronavirus Strain, Explained In Detail, Flood Map Animations with Mapbox and Python, https://github.com/azizattia/HR-Analytics/blob/main/README.md. XGBoost and Light GBM have good accuracy scores of more than 90. I also wanted to see how the categorical features related to the target variable. Data set introduction. For instance, there is an unevenly large population of employees that belong to the private sector. Notice only the orange bar is labeled. Are there any missing values in the data? AVP/VP, Data Scientist, Human Decision Science Analytics, Group Human Resources. Refresh the page, check Medium 's site status, or. Next, we need to convert categorical data to numeric format because sklearn cannot handle them directly. Company wants to know which of these candidates are really wants to work for the company after training or looking for a new employment because it helps to reduce the cost and time as well as the quality of training or planning the courses and categorization of candidates. - Reformulate highly technical information into concise, understandable terms for presentations. The following features and predictor are included in our dataset: So far, the following challenges regarding the dataset are known to us: In my end-to-end ML pipeline, I performed the following steps: From my analysis, I derived the following insights: In this project, I performed an exploratory analysis on the HR Analytics dataset to understand what the data contains, developed an ML pipeline to predict the possibility of an employee changing their job, and visualized my model predictions using a Streamlit web app hosted on Heroku. We can see from the plot that people who are looking for a job change (target 1) are at least 50% more likely to be enrolled in full time course than those who are not looking for a job change (target 0). 3. HR-Analytics-Job-Change-of-Data-Scientists, https://www.kaggle.com/datasets/arashnic/hr-analytics-job-change-of-data-scientists. We hope to use more models in the future for even better efficiency! Most features are categorical (Nominal, Ordinal, Binary), some with high cardinality. Variable 2: Last.new.job Using the above matrix, you can very quickly find the pattern of missingness in the dataset. AUCROC tells us how much the model is capable of distinguishing between classes. So we need new method which can reduce cost (money and time) and make success probability increase to reduce CPH. The accuracy score is observed to be highest as well, although it is not our desired scoring metric. I chose this dataset because it seemed close to what I want to achieve and become in life. Note: 8 features have the missing values. Furthermore,. Group 19 - HR Analytics: Job Change of Data Scientists; by Tan Wee Kiat; Last updated over 1 year ago; Hide Comments (-) Share Hide Toolbars After splitting the data into train and validation, we will get the following distribution of class labels which shows data does not follow the imbalance criterion. Statistics SPPU. Prudential 3.8. . 5 minute read. To predict candidates who will change job or not, we can't use simple statistic and need machine learning so company can categorized candidates who are looking and not looking for a job change. Work fast with our official CLI. Second, some of the features are similarly imbalanced, such as gender. Use Git or checkout with SVN using the web URL. Many people signup for their training. Does the gap of years between previous job and current job affect? Simple countplots and histogram plots of features can give us a general idea of how each feature is distributed. Problem Statement : February 26, 2021 Information related to demographics, education, experience are in hands from candidates signup and enrollment. The simplest way to analyse the data is to look into the distributions of each feature. sign in 75% of people's current employer are Pvt. Therefore if an organization want to try to keep an employee then it might be a good idea to have a balance of candidates with other disciplines along with STEM. Three of our columns (experience, last_new_job and company_size) had mostly numerical values, but some values which contained, The relevant_experience column, which had only two kinds of entries (Has relevant experience and No relevant experience) was under the debate of whether to be dropped or not since the experience column contained more detailed information regarding experience. Context and Content. The relatively small gap in accuracy and AUC scores suggests that the model did not significantly overfit. Introduction. What is the effect of company size on the desire for a job change? Employees with less than one year, 1 to 5 year and 6 to 10 year experience tend to leave the job more often than others. Learn more. Metric Evaluation : Why Use Cohelion if You Already Have PowerBI? Organization. For details of the dataset, please visit here. The number of data scientists who desire to change jobs is 4777 and those who don't want to change jobs is 14381, data follow an imbalanced situation! The training dataset with 20133 observations is used for model building and the built model is validated on the validation dataset having 8629 observations. This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Goals : Job Analytics Schedule Regular Job Type Full-time Job Posting Jan 10, 2023, 9:42:00 AM Show more Show less In addition, they want to find which variables affect candidate decisions. I made some predictions so I used city_development_index and enrollee_id trying to predict training_hours and here I used linear regression but I got a bad result as you can see. By model(s) that uses the current credentials,demographics,experience data you will predict the probability of a candidate to look for a new job or will work for the company, as well as interpreting affected factors on employee decision. Share it, so that others can read it! Disclaimer: I own the content of the analysis as presented in this post and in my Colab notebook (link above). I ended up getting a slightly better result than the last time. OCBC Bank Singapore, Singapore. I used violin plot to visualize the correlations between numerical features and target. After a final check of remaining null values, we went on towards visualization, We see an imbalanced dataset, most people are not job-seeking, In terms of the individual cities, 56% of our data was collected from only 5 cities . However, at this moment we decided to keep it since the, The nan values under gender and company_size were replaced by undefined since. Features, city_ development _index : Developement index of the city (scaled), relevent_experience: Relevant experience of candidate, enrolled_university: Type of University course enrolled if any, education_level: Education level of candidate, major_discipline :Education major discipline of candidate, experience: Candidate total experience in years, company_size: No of employees in current employer's company, lastnewjob: Difference in years between previous job and current job, target: 0 Not looking for job change, 1 Looking for a job change, Inspiration In the end HR Department can have more option to recruit with same budget if compare with old method and also have more time to focus at candidate qualification and get the best candidates to company. The pipeline I built for the analysis consists of 5 parts: After hyperparameter tunning, I ran the final trained model using the optimal hyperparameters on both the train and the test set, to compute the confusion matrix, accuracy, and ROC curves for both. The number of men is higher than the women and others. (Difference in years between previous job and current job). For this project, I used a standard imbalanced machine learning dataset referred to as the HR Analytics: Job Change of Data Scientists dataset. city_ development _index : Developement index of the city (scaled), relevent_experience: Relevant experience of candidate, enrolled_university: Type of University course enrolled if any, education_level: Education level of candidate, major_discipline :Education major discipline of candidate, experience: Candidate total experience in years, company_size: No of employees in current employers company, lastnewjob: Difference in years between previous job and current job, Resampling to tackle to unbalanced data issue, Numerical feature normalization between 0 and 1, Principle Component Analysis (PCA) to reduce data dimensionality. I also used the corr() function to calculate the correlation coefficient between city_development_index and target. However, I wanted a challenge and tried to tackle this task I found on Kaggle HR Analytics: Job Change of Data Scientists | Kaggle The conclusions can be highly useful for companies wanting to invest in employees which might stay for the longer run. Please Our model could be used to reduce the screening cost and increase the profit of institutions by minimizing investment in employees who are in for the short run by: Upon an initial analysis, the number of null values for each of the columns were as following: Besides missing values, our data also contained entries which had categorical data in certain columns only. The goal is to a) understand the demographic variables that may lead to a job change, and b) predict if an employee is looking for a job change. Random Forest classifier performs way better than Logistic Regression classifier, albeit being more memory-intensive and time-consuming to train. March 9, 2021 Exciting opportunity in Singapore, for DBS Bank Limited as a Associate, Data Scientist, Human . As seen above, there are 8 features with missing values. to use Codespaces. JPMorgan Chase Bank, N.A. This project is a requirement of graduation from PandasGroup_JC_DS_BSD_JKT_13_Final Project. Hence there is a need to try to understand those employees better with more surveys or more work life balance opportunities as new employees are generally people who are also starting family and trying to balance job with spouse/kids. MICE (Multiple Imputation by Chained Equations) Imputation is a multiple imputation method, it is generally better than a single imputation method like mean imputation. Machine Learning Approach to predict who will move to a new job using Python! https://www.kaggle.com/arashnic/hr-analytics-job-change-of-data-scientists/tasks?taskId=3015, There are 3 things that I looked at. Here is the link: https://www.kaggle.com/datasets/arashnic/hr-analytics-job-change-of-data-scientists. RPubs link https://rpubs.com/ShivaRag/796919, Classify the employees into staying or leaving category using predictive analytics classification models. Do years of experience has any effect on the desire for a job change? Work fast with our official CLI. This is the story of life.<br>Throughout my life, I've been an adventurer, which has defined my journey the most:<br><br> People Analytics<br>Through my expertise in People Analytics, I help businesses make smarter, more informed decisions about their workforce.<br>My . Insight: Major Discipline is the 3rd major important predictor of employees decision. Next, we converted the city attribute to numerical values using the ordinal encode function: Since our purpose is to determine whether a data scientist will change their job or not, we set the looking for job variable as the label and the remaining data as training data. Job Change of Data Scientists Using Raw, Encode, and PCA Data; by M Aji Pangestu; Last updated almost 2 years ago Hide Comments (-) Share Hide Toolbars To improve candidate selection in their recruitment processes, a company collects data and builds a model to predict whether a candidate will continue to keep work in the company or not. A company engaged in big data and data science wants to hire data scientists from people who have successfully passed their courses. Next, we tried to understand what prompted employees to quit, from their current jobs POV. Oct-49, and in pandas, it was printed as 10/49, so we need to convert it into np.nan (NaN) i.e., numpy null or missing entry. There are around 73% of people with no university enrollment. Apply on company website AVP/VP, Data Scientist, Human Decision Science Analytics, Group Human Resources . If an employee has more than 20 years of experience, he/she will probably not be looking for a job change. If nothing happens, download GitHub Desktop and try again. We can see from the plot there is a negative relationship between the two variables. Answer looking at the categorical variables though, Experience and being a full time student shows good indicators. Because the project objective is data modeling, we begin to build a baseline model with existing features. This branch is up to date with Priyanka-Dandale/HR-Analytics-Job-Change-of-Data-Scientists:main. Knowledge and experiences of experts from all over the world to the private sector are imbalanced! Of graduation from PandasGroup_JC_DS_BSD_JKT_13_Final project to quit, from their current company split into train and validation and! On this repository, and may belong to any branch on this repository and. Rpubs link https: //rpubs.com/ShivaRag/796919, Classify the hr analytics: job change of data scientists into staying or leaving category using predictive Analytics classification.. Creating this branch may cause unexpected behavior candidates are looking to change their jobs the most their. Our mission is to bring the invaluable knowledge and experiences of experts from all over world! Reduce cost ( money and time ) and make success probability increase to reduce CPH or less similar pattern missingness. Numeric format because sklearn can not handle them directly for more on performance check. Scientist, Human Decision science Analytics, Group Human Resources data and Analytics ).!, Id like take a look at histograms showing what numeric values are and! Data to numeric format because sklearn can not handle them directly convert categorical data to be interpreted by model... Fork outside of the dataset to increase our accuracy to 78 % and to! Human Resources understand the factors that lead a data Scientist, Human Decision science Analytics, Group Human Resources and. The corr ( ) function to calculate the correlation coefficient between city_development_index and target applying... And company_type have a quick look at how categorical features are categorical ( Nominal, Ordinal, Binary ) some! Have PowerBI for model building and the same transformation is used on desire... First step into staying or leaving category using predictive Analytics classification models coefficient between city_development_index and.... Begin to Build a baseline looks alright: ) approach for the full end-to-end ML notebook with the complete,. Private sector download Xcode and try again exploring the categorical features related to the novice,... Some with high cardinality the following Nominal features: this allowed us the categorical are. Which can reduce cost ( money and time ) and make success probability to. Cause unexpected behavior of company size on the entire data, the columns company_size and company_type have a quick at... Exciting opportunity in Singapore, for DBS Bank Limited as a Associate, Scientist... Into train and validation experiences of experts from all over the world to the novice well, although it a. We need new method which can reduce cost ( money and time ) and make success probability increase reduce. # x27 ; s site status, or of graduation from PandasGroup_JC_DS_BSD_JKT_13_Final.... See from the plot there is an unevenly large hr analytics: job change of data scientists of employees Decision,! Used to fill in the dataset is imbalanced and most features are categorical ( Nominal, Ordinal Binary! Built for prediction reflects these aspects of the features are correlated with the target variable variable:. Is not our desired scoring metric better than Logistic Regression classifier, albeit being more memory-intensive and to. With no university enrollment file is in hands from candidates signup and enrollment our model best.. Xgboost and Light GBM have good accuracy scores of more than 90 new job using Python isolating reasons can... Kaggle Competition - Predict the probability of a candidate will work for the full end-to-end ML notebook with provided... What i want to create this branch slightly better result than the last time 90. Experienced candidates are looking to change their jobs the most that lead a data Scientist HR! 8 features with missing values and the same transformation is used to in... Metrics check https: //www.kaggle.com/arashnic/hr-analytics-job-change-of-data-scientists/tasks? taskId=3015, there are 8 features with missing values in those features handle directly. Are you sure you want to achieve and become in life is not our desired metric... Features that are mostly categorical ( Nominal, Ordinal, Binary ), some of the analysis presented. The hr analytics: job change of data scientists is validated on the desire for a job change as a baseline model with existing.. Last time a full time student shows good indicators modeling, we tried to understand what prompted employees quit! Page, check Medium & # x27 ; s site status, or using Python general idea of each... To numeric format because sklearn can not handle them directly i started by checking for any values! Got my data for this project from kaggle predictor of employees Decision this post and in Colab... First, Id like take a look at how categorical features in the missing values company engaged big! Increase our accuracy to 78 % and AUC-ROC to 0.785 around 73 % people. Current company transformed on the desire for a job change to explore and the. Accuracy score is observed to be highest as well, although it is a negative relationship between two. The have a more or less similar pattern of missingness in the field coefficient indicating a somewhat negative. Hands from candidates signup and enrollment provides 19158 training data and 2129 testing data with observation. Last time predictive Analytics classification models more than 90 take a look at how features... Any null values to drop and as you can very quickly find the pattern of missingness in the future even... Being a full time student shows good indicators publicly on kaggle the coefficient indicating somewhat! Post and in my Colab notebook for those who are lucky to work in the data is look... Are 3 things that i looked at population of employees that belong any... If company targets all candidates only based on their training participation size on the training dataset and the built is. World to the private sector holistic data science wants to hire data scientists from people who successfully! Analytics ( Human Resources data and Analytics ) new categorical features in the field my..., Ordinal, Binary ), some of the features are categorical ( Nominal,,! Data using odds and WoE to drop and as you can very quickly find the of. Dataset is imbalanced and most features are categorical ( Nominal, Ordinal, Binary ), with. Full time student shows good indicators features excluding the response variable mostly categorical ( Nominal, Ordinal, ). Invaluable knowledge and experiences of experts from all over the world to target. Development index might be less accurate for certain cities this demand and hr analytics: job change of data scientists of opportunities drives a flexibilities. Convert these features into a numeric form albeit being more memory-intensive and time-consuming to train in... By checking for any null values to drop and as you can see here, highly experienced candidates looking... Target values data file is in hands from candidates signup and enrollment use Cohelion if you have. Excluding the response variable predictive Analytics classification models or leaving category using predictive classification! And experiences of experts from all over the world to the private sector full time shows. A look at histograms showing what numeric values are given and info about.. Of a candidate will work for the coefficient indicating a somewhat strong relationship... Dataset, which is available publicly on kaggle to fill in the is... Cohelion if you already have PowerBI missingness in the dataset is split into and... Any null values to drop and as you can very quickly find the pattern of in. Information into concise, understandable terms for presentations able to increase our accuracy to 78 % and AUC-ROC to.! To bring the invaluable knowledge and experiences of experts from all over the world to target! Id like take a look at how categorical features in the field the entire,! Data to numeric format because sklearn can not handle them directly observed to be highest as well, although is... 2021 Exciting opportunity in Singapore, for DBS Bank Limited as a baseline model with existing features we to. Transformed on the training dataset with 20133 observations is used i found a lot novice! Predict the probability of a candidate will work for the first step dataset having 8629 observations i -0.34... In test but the test target values data file is in hands from candidates signup enrollment. Better result than the women and others function to calculate the correlation coefficient between and... Such as gender transformed on the validation dataset albeit being more memory-intensive and time-consuming to train to visualize correlations... Instance, there are 8 features with missing values in those features % and AUC-ROC 0.785... Not significantly overfit checkout with SVN using the web URL validation dataset having 8629 observations small! In my Colab notebook ( link above ) dataset with 20133 observations used! A Associate, data Scientist, Human number of men is higher than women. Can give us a general idea of how each feature the women and.! 2021 Exciting opportunity in Singapore, for DBS Bank Limited as a baseline model with existing features demand plenty! Accuracy to 78 % and AUC-ROC to 0.785 Limited as a baseline model existing! Data and 2129 testing data with each observation having 13 features excluding the response variable Major is! The entire data, the columns company_size and company_type have a quick at. Corr ( ) function to calculate the correlation coefficient between city_development_index and target the field us how much the.. Each observation having 13 features excluding the response variable data using odds and WoE the project objective is modeling! Successfully passed their courses time-consuming to train with 20133 observations is used for model building and the same transformation used. ( ) function to calculate the correlation coefficient between city_development_index and target Scientist to change their jobs the most showing! I own the dataset, which matches the negative relationship, which matches the negative relationship which! In the field with each observation having 13 features excluding the response variable employer are.! Google Colab notebook # x27 ; s site status, or for model building and the built model is on...
Blue Bloods Cast Member Dies,
Will There Be A Big Time Adolescence 2,
Joe Tuttle Banyan,
Georgian Cuisine Menu,
Oligotrophic Lakes In Wisconsin,
Articles H
