Luna Baalbaki - IBM HR Employee Turnover - Datathon LSE

Hello :) I am Luna Baalbaki and here's some information about my Datathon project.

Inspiration

I have seen on hand how data is powerful. I enjoy using data to inspire decision makers.

What it does

IBM is struggling from a 16% employee attrition rate. But luckily they have great data that we can help to analyze. I used data analysis and machine learning models to clearly point out to a few of the top reasons and signals that contribute to the high rate of employee attrition. I suggest at the end some recommendations to help the HR department develop strategies to decrease the employee attrition rate.

How we built it

I have used Jupyter Notebook and python code to work on the data. I have done:

• Data cleaning
• Explatory Analysis _ Feature Engineering
• One Hot Encoding
• Machine Learning Model Selection
• Parameter Optimization
• Feature Selection

Challenges we ran into

I have faced some challenges in the machine learning model selection. I have realized that some of the columns have textual data in them (they are strings) and that they don't work with machine learning models. I have done some research yesterday night and I have realized that I have to do one hot encoding such that I change all the textual data into numbers 1-0 so that they could work with machine learning models. Then that's what I did and it worked.

Accomplishments that we're proud of

I am proud that my results and insights are logincal and make sense. I am proud of myself a lot because I am not used to data science a lot.

What we learned

I have learnt many things from this Datathon. Firstly the workshops were very helpful. Secondly, while working on the project I researched a lot of things and learnt many new information such as feature engineering and so on.

What's next for Luna Baalbaki - IBM HR Employee Turnover - Datathon LSE

I am planning to build web app for the analysis that I have built. That would be on Streamlit. Where the company would enter the data of a new employee and then the machine learning model that I have chosen (based on accuracy and better prediction) would suggest the percentage of that employee leaving the company. From here the HR department can work on strategies to decrease the employee attrition rate.

Built With

• jupyternotebook
• python