Inspiration
According to an online article "The Bureau of Labor Statistics (BLS) reported that 3.5 million people—or 2.3 percent of the total workforce—left their jobs voluntarily in October, the most recent month for which data exists. Private-sector workers quit at a rate of 2.6 percent in October, up from 2.4 percent the previous year." In our current growing technology needs and making organization at top of the world, there is one thing most of the organizations missing which is work engagement. Work engagement is simply described as the active involvement of a person towards their work. We strongly believe that measuring the work engagement through the available data of a person within an organization can lead to interesting root cause analysis. Further, it can be used to help take the necessary action to make a person's work engagement even better.
What it does
Our application takes necessarily a wide variety of data related to a person in an organization. The data could be such as working hours, clocked in and clocked out time, sentimental analysis from communication channels, salary hike, distance from home and others. The application runs an engine where all the data is fed and it will predict the work engagement measure on four levels. It is up to the organization to decide the threshold for low work engagement based on their needs.
How I built it
We built primarily using data available from the Kaggle website (https://www.kaggle.com/vjchoudhary7/hr-analytics-case-study). We created our engine in python using powerful computing and deep learning libraries such as Pandas, Numpy, SkLearn, PyTorch. In the end, our application runs on AWS for more scalability and getting quick insights through the dashboard.
Challenges I ran into
Data is the new oil. Just like how extracting oil is a difficult process, processing and harvesting the right data specific to a certain problem is a tough process. We spend half of the time in collecting the data and integrating it to make it cohesive. Also building the dashboard and making it linked with our engine is also a challenging task.
Accomplishments that I'm proud of
We are proud of tackling a problem that is pervasive in our current society. This domain is entirely new for us and there is a lot to explore. We felt that we have accomplished a tiny step in this area where measuring work engagement can lead to proactively develop solutions that many organizations facing.
What I learned
Data is ubiquitous. All we need is to take a look at the bigger picture and start from the lowest possible grain to solve open-ended problems.
What's next for WEMINO: Work Engagement Measure IN Organizations
The next step is to make our engine more robust by adding even more factors that can contribute to a person's work engagement. A further step could be to develop a Natural Language Processing model that can directly give suggestions to both workers and the organizations on how to improve work engagement.
Log in or sign up for Devpost to join the conversation.