Salary.io

Motivation

Whether you're job searching or have an offer in hand, knowing what salary you should be paid is important. Our tool takes into account where you live, your title, and how many years you have been working to estimate a salary that you can use as a reference throughout your career journey.

What it does

Our app allows you to enter some information such as years of experience, gender and location about yourself and it will predict your salary based on a linear regression model trained by real world data.

Challenges we ran into

None of us have much experience in machine learning and we spent a lot of time trying to identify the best dataset and which model to use. We used salary data from Levels.fyi. It was difficult to train a model using user-reported data, which contains a lot of noise and inaccuracies that led to a lot of variance. We trained both a sklearn linear regression model and a keras DNN model but their performance was similar.

Accomplishments that we're Proud of

We were able to produce a finished product that utilizes either model for our first-time machine learning project!

What We Learned

How much time we need to invest in looking for datasets and different model types.

What's next for Salary.io

The ability to compare yourself against others in the market on a graph would be awesome!

Technologies Used :rocket:

Built with

• Language: Python, Javascript
• Flask
• React
• sklearn

Made with :heart: by GitGeese (Pod 1.1.6)

| Lily Tang | Gerald Marquez | Prionti Nasir | Shiyue Zhang |

Built With

• flask
• javascript
• python
• react
• scikit-learn