Determine the dynamics of which features predict New York City Airbnb rental prices.
Develop an algorithm to accurately predict prices using those features.
Training dataset and environment.yml (recipe to create a new conda env) attached below.
Submissions will be graded on model accuracy, performance and clarity of code.
Note: We are using an open kaggle dataset that has Airbnb listings in New York and their associated prices, sourced from insideairbnb.com
- Download the environment recipe below (versions are specified in the file)
stemaway_ML_env.yml (1.2 KB)
NOTE: Please remove ‘glmnet’ if this library is causing an issue. You can create a submission without the help of this library.
- To build the environment, you need anaconda or miniconda
- Execute the steps below to ensure your submission can be checked. Executing these steps also ensures that everyone engineers their submission with identical environments.
conda env create -f stemaway_ML_env.yml -> This will attempt to build the environment. It may outright fail, or fail building a particular package and build the env up to the package conda activate stemaway_ML_env -> If the above step works, the environment was setup properly conda list -> This will show the packages involved
Go to STEM-Away® Machine Learning Evaluation to submit the following files (you will need to create a zip of the files)
- train.py - code used to train and save your model
- predict.py - defines a function called ‘predict.py’ that takes as input test data and returns your model’s predictions for that test data
- model.json (or any other serialized file type) - the actual model itself that gets loaded in predict.py
Please visit the recordings under Discussion Board to learn more about the setup, goals, possible algorithms and submission templates.
- Your first submission is the one that is graded. If you are not sure about the material, please wait until the corresponding training is held.
- Highest level of academic integrity is expected from all members. Subtle changes are made in the test to discourage copying of answers.
Colin Magdamo, STEM-Away® Principal Mentor