Mental Health Stats

Predicts the status of mental health attitude of the user in the work place, and provides recommendations accordingly

Comment

Inspiration

We wanted to work towards building a recommendation engine or similar tool that addresses and raises awareness of mental health issues in tech.

What it does

It asks for the user to input some factors like age, gender, type of company, state etc and predicts the possible outcome for the mental health attitude with the help of training data provided by kaggle.com.

How we built it

We first used Adobe XD to create a prototype for our web app. We got the training data for mental health in tech from Kaggle.com. Then we used Python to explore relationships between data features, to analyze and visualize the data. We used Anaconda, Pandas and most importantly the scikit library to find the logistic, linear regression, barplot, and Pie-chart. We Used ChartJS to visualize/highlight key findings.

Challenges we ran into

According to our mentor, one of the major problems in the world of data science is invalid entry/typos in the input. Due to that, we were missing out a huge number of data sets that were negatively affecting our prediction. Thus, we had to go back and clean our data. We discarded the invalid input and fixed some typos ourselves and simplified the categories. Another challenge that my teammate Oriane ran into was to read and process data from CSV file with JavaScript and cleaning the data.

Accomplishments that we're proud of

We both had different ideas on what we wanted to do for the hackathon. But we were able to combine our interests and skill sets to create something. I was more interested in Machine learning, and my teammate Oriane was interested in interactive data visualization so that it would be easier for the screen reader to actually describe the data. So, we both worked on what we wanted to work on even if we didn't have a set outcome. Later on, everything just falled on place.

What we learned

One of the most important things we learned was to play with data before doing anything. We realized that analyzing data at first is important before trying to jump to a conclusion. Also, we also learned to not hesitate to ask for help to mentors. 1:1 can be really effective when you're stuck at something for long.

What's next for Mental Health in Tech

With more key findings, we can begin to better predict what mental health related information/resources/action might be most useful for an individual based on certain factors.

Built With

Share this project:

Updates