Two Sigma Connect: Rental Listing Inquiries

Kaggle Competition (link)

Create Kaggle account

Join competition

Mount gdrive

from google.colab import drive
drive.mount('/content/drive')

Import / Config

import pandas as pd
import os

kaggle_dir = f"/content/drive/My Drive/kaggle"
competition = "two-sigma-connect-rental-listing-inquiries"
target_dir = f"{kaggle_dir}/rent-two-sigma"

!mkdir -p  "{kaggle_dir}"

os.environ['KAGGLE_CONFIG_DIR'] = kaggle_dir

Data Preparation

only done once
skip this cell if you start your notebook anew
Open Kaggle API doc (link)
Goto section authentication
Create kaggle.json
Upload kaggle.json to kaggle directory on your gdrive

!kaggle competitions download -c "{competition}"
!unzip {competition}.zip -d "{target_dir}"
!unzip "{target_dir}/train.json.zip" -d "{target_dir}"

df = pd.read_json(f"{target_dir}/rent-two-sigma/train.json")
df.to_parquet(f"{target_dir}/rent.parquet")

df = df[(df.price>1_000) & (df.price<10_000)]
df = df[(df.longitude!=0) | (df.latitude!=0)]
df = df[(df['latitude']>40.55) & (df['latitude']<40.94) &
        (df['longitude']>-74.1) & (df['longitude']<-73.67)]
df_num = df[['bedrooms','bathrooms','latitude','longitude','price']]
df_num.to_parquet(f"{target_dir}/rent-ideal.parquet")

!rm "{target_dir}/images_sample.zip"
!rm "{target_dir}/Kaggle-renthop.torrent"
!rm "{target_dir}/sample_submission.csv.zip"
!rm "{target_dir}/test.json.zip"
!rm "{target_dir}/train.json"
!rm "{target_dir}/train.json.zip"

Training a random forest model

Link to ML book at explained.ai (link)

rent = pd.read_parquet(f"{target_dir}/rent-ideal.parquet")
rent

Exploring and Denoising Your Data Set

Link to ML book at explained.ai (link)

df = pd.read_parquet(f"{target_dir}/rent.parquet")
df

Categorically Speaking

Link to ML book at explained.ai (link)

df = pd.read_parquet(f"{target_dir}/rent.parquet")
df