Autolib-Project

Problem statement

To work as a Data Scientist for the Autolib electric car-sharing service company to investigate a claim about the blue cars from the Autolib dataset on either weekday or weekend.

Creating Hypothesis

H0:Bluecars are mostly taken on weekdays.

H1: Bluecars are not mostly taken on weekdays.

Importing libaries and Loading the datasets

import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns from scipy import stats from statsmodels.stats import weightstats from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler from sklearn.metrics import confusion_matrix,accuracy_score from sklearn.naive_bayes import GaussianNB

Data Understanding

it has 'Postal code', 'date', 'n_daily_data_points', 'dayOfWeek', 'day_type','BlueCars_taken_sum', 'BlueCars_returned_sum', 'Utilib_taken_sum', 'Utilib_returned_sum', 'Utilib_14_taken_sum', 'Utilib_14_returned_sum', 'Slots_freed_sum', 'Slots_taken_sum' columns

It has 4645 rows

Data Cleaning

Checking for outliers

Data Visualization Checking for varibles that correlate uasing scatter plot and heatmap

checking for the frequency of variables using bar graph pie chart and histogram

Use of sampling techniques

simple random sampling

stratified random sampling

Testing The significance of a test

Tested normality using qqlot and histograms

using z_test to check for significance of the hypothesis

From the above analysis we conclude that usage of bluecars is more on weekdays compared to weekend

lydiamokaya/autolib