Data Science Portfolio
Here is a list of projects related to data science that I have done.
- ConsciousChef: A recipe recommender app that I created as a part of the insight data science fellowship program. This app is targeted towards CSA users, whose major concern is to use up all the produce in the CSA box and to reduce wastage. This app recommends recipes by optimizing for shelf life and also the quantity of produce. For this project I scraped the recipe information from epicurious.com. You can read about this more here. You can find the github repo here.
Restaurant Recommender App: I wanted to get practice dealing with web scraping and also working with yelp and google API’s. I had it running with heroku apps. But keeping it up-to-date was a challenge. So I let it die. Here is the link to git repo.
Machine Learning Algorithms Practice: Here I wanted to practice my ML skills while I prepared for interviews, so I used the well known and interesting Kaggle and UCI datasets to solve problems.
- Is the tumor malignant or benign?: The goal of this exercise was to predict if a tumor was benign or malignant. This was a great dataset to look into various classification approaches.
** Algorithms used: PCA, t-SNE, Logistic Regression, Support Vector Machines, Random Forests, KNN
- House price prediction: This dataset was a great intro to regression algorithms.
** Algorithms used: Linear Regression (Lasso, Ridge, ElasticNet), Random Forests, Gradient Boosting
- Smartphone activity prediction: This dataset contains the mobile phone accelerometer data collected from 30 subjects. The dataset is labelled with the subjects activity name (STANDING, WALKING etc).
** Algorithms used: PCA, KNN, Random Forests
- Whale sound detection: It was easy to relate to this dataset because I have been working in bioacoustics for the entirety of my research career.
whale call not a whale call
Algorithms used: FFT, PCA, Random Forests