Skip to content

Instantly share code, notes, and snippets.

View djsegal's full-sized avatar

djsegal

83 followers · 273 following
  • home
View GitHub Profile
@djsegal
djsegal / ensemble_pipeline.py
Created January 5, 2020 00:26
cur_data = price_data.copy()
cur_data = cur_data.merge(rental_data, on="id")
cur_data = cur_data.merge(location_data, on="id")
cur_data = cur_data.merge(geo_data, on="id")
X_train, X_test, y_train, y_test = \
custom_train_test_split(cur_data)
pass_cols = ["is_brooklyn", "density"]
drop_cols = ["year", "geometry", "zipcode"]
@djsegal
djsegal / model_feature_importances.csv
Created January 3, 2020 07:05
accommodates accommodates _per_bathrooms accommodates _per_bedrooms accommodates bedrooms center_dist _is_brooklyn food_density food_density _is_brooklyn is_brooklyn
All NYC 34 -14 -9 -4 -6 8 16 -9
Manhattan 51 -15 -13 -13 - 9 - -
Brooklyn 43 -19 -16 -7 - 16 - -
Linear SVM 25 -11 -7 -6 -1 3 21 -25
XGBoost 7 1 1 38 29 22 1 0
AdaBoost 4 1 0 25 14 51 4 0
Extra Forest 6 4 4 8 16 45 11 7
Random Forest 2 1 2 26 11 56 3 0
@djsegal
djsegal / step_4_3.csv
Created January 3, 2020 03:48
Feature Relative Importance
accommodates 1.0
accommodates _per_bathrooms -0.428
accommodates _per_bedrooms -0.28
food_density 0.248
is_brooklyn -0.231
center_dist -0.102
density -0.086
@djsegal
djsegal / nyc_airbnb_magnitudes.csv
Created January 1, 2020 08:21
Dataset Count Parameter Descriptor
Apartments 15k N Airbnb
Restaurants 15k M DOH
Subway Stations 300 M MTA
Hotspots 10 M Red Diamonds
Cost Epicenter 1 M Orange Star
@djsegal
djsegal / new_york.geojson
Last active January 1, 2020 07:18
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@djsegal
djsegal / cluster_data.py
Created January 1, 2020 02:18
import numpy as np
import pandas as pd
import geopandas
import pickle
from sklearn.cluster import KMeans
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
@djsegal
djsegal / cluster_data.csv
Last active January 4, 2020 05:42
name lat lon price is_com is_bkn
0 center 40.7269845195146 -73.96825439838058 170 True False
1 wash sq 40.73232616009288 -74.00113694319073 197 False False
2 flatiron 40.74010002079383 -73.99202965334189 555 False False
3 bowery 40.728100910208695 -73.99392182445409 298 False False
4 uptown 40.79674265987271 -73.95298422453986 302 False False
5 midtown 40.76134044954954 -73.98369000821411 303 False False
6 barclays 40.68061129890911 -73.97564856875336 267 False True
7 bushwick 40.688196501332044 -73.92715263054822 148 False True
8 williamsburg 40.715948534974885 -73.95220786983514 225 False True
@djsegal
djsegal / step_4_1.csv
Created December 30, 2019 08:04
feature coefficient
accommodates 1.0
accommodates _per_bathrooms -0.43
is_brooklyn -0.405
accommodates _per_bedrooms -0.338
accommodates bedrooms -0.305
density -0.115
accommodates bathrooms 0.111
bedrooms _per_beds -0.108
accommodates _per_beds 0.105
@djsegal
djsegal / step_4_1.py
Last active January 1, 2020 04:34
cur_data = price_data.copy()
cur_data = cur_data.merge(rental_data, on="id")
cur_data = cur_data.merge(location_data, on="id")
X_train, X_test, y_train, y_test = \
custom_train_test_split(cur_data)
pass_cols = ["is_brooklyn", "density"]
drop_cols = ["year", "geometry", "zipcode"]
one_hot_cols = ["month"]
@djsegal
djsegal / pass_through_transformer.py
Created December 30, 2019 04:16
from sklearn.base import BaseEstimator, TransformerMixin
class PassThroughTransformer(BaseEstimator, TransformerMixin):
def __init__(self):
self.input_features = None
def fit(self, X, y=None):
assert self.input_features is None
if type(X) == np.ndarray :