#Uber Visualization Uber V/S Lyft

Your project is for visulizing the data of uber and lyft comparision.This data is of UK

Code :

#import the libary

import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns from datetime import datetime

Load the data

cab_df = pd.read_csv("cab_rides.csv")

#It the header of data #In the we can see the column of data cab_df.head()

#checking the null value cab_df.isnull().sum()

We see that price column has 55095 rows missing data.

#Total rows we have 693071 so we can remove the rows with no price. cab_df.dropna(inplace=True) #Againg seeing that null value cab_df.isnull().sum()

First, lets see what columns do we have

cab_df.columns

We don't need id product_id surge_multiplier. So lets get rid of them

cab_df = cab_df[['distance', 'cab_type', 'time_stamp', 'destination', 'source', 'price', 'name']] cab_df.head()

We see that time_stamp is Unix, so we need to convert it to the readable form.

Using the time_stamp column, lets convert it to date, week day, hour and time of day.

cab_df["rounded_timestamp"] = cab_df["time_stamp"] / 1000 cab_df["rounded_timestamp"] = cab_df["rounded_timestamp"].apply(np.floor)

#we are creating four different column cab_df["date"] = cab_df["rounded_timestamp"].apply(lambda x : datetime.fromtimestamp(x).date()) cab_df["time"] = cab_df["rounded_timestamp"].apply(lambda x: datetime.fromtimestamp(x).time()) cab_df['weekday'] = cab_df['date'].apply(lambda x: x.weekday()) cab_df["weekday"] = cab_df["weekday"].map({0: 'Monday', 1: 'Tuesday', 2: 'Wednesday', 3: 'Thursday', 4: 'Friday', 5: 'Saturday', 6: 'Sunday'}) cab_df['hour'] = cab_df['time'].apply(lambda time: time.hour)

#see the date cab_df['date'].head()

#see the time cab_df['time'].head()

We calculate time of day into: Morning,Afternoon,Evening and Night

cab_df.loc[(cab_df.hour >= 6) & (cab_df.hour < 12) , 'time_of_day'] = 'Morning' cab_df.loc[(cab_df.hour >= 12) & (cab_df.hour < 16) , 'time_of_day'] = 'Afternoon' cab_df.loc[(cab_df.hour >= 16) & (cab_df.hour < 22) , 'time_of_day'] = 'Evening' cab_df.loc[(cab_df.hour >= 22) | (cab_df.hour < 6) , 'time_of_day'] = 'Night'

#After adding column 'date', 'time','weekday', 'hour', 'time_of_day' see that cab_df.columns

#need of column which is required for uber visualization cab_df = cab_df[['distance', 'cab_type', 'time_stamp', 'price', 'name', 'date', 'time', 'weekday', 'hour', 'time_of_day']]

#if you want this new data set then use with conver of date and time from unix format cab_df.to_csv("new.csv") cab_df.head()

cab_df['cab_type'].value_counts()

#see the sift of time cab_df['time_of_day'].value_counts()

So we can see we have two cab types: Uber and Lyft

So we need to separate the datasets

uber_df = cab_df[cab_df['cab_type'] =="Uber"] lyft_df = cab_df[cab_df['cab_type'] =="Lyft"]

From above we see that for Uber we have only one value of price

So we consider the price for uber only and plot which day it is highest.

We only consider the price > 1.

high_price_dataset = uber_df[uber_df["price"]>1] high_distance_dataset = uber_df[uber_df["distance"]> 0.1]# From above we see that for Uber we have only one value of price

So we consider the price for uber only and plot which day it is highest.

We only consider the price > 1.

high_price_dataset = uber_df[uber_df["price"]>1] high_distance_dataset = uber_df[uber_df["distance"]> 0.1]

t_high_price = pd.DataFrame(high_price_dataset.groupby(["weekday", "price"]).size().reset_index()) t_high_price.columns = ["Weekday", "price","hour"] plt.figure(figsize=(15, 6)) sns.barplot(x="Weekday", y="price", data=t_high_price,ci=None,estimator=np.max).set_title("Weekday wise price range");

t_high_price.groupby("Weekday").price.describe()

t_high_price1 = pd.DataFrame(high_price_dataset.groupby(["weekday", "price","name"]).size().reset_index()) t_high_price1.columns = ["Weekday", "price","name","count"] plt.figure(figsize=(22, 5)) sns.barplot(x="Weekday", y="count", hue="name", data=t_high_price1,ci=None,estimator=np.mean).set_title("Weekday wise Surge");

td_high_day = pd.DataFrame(high_distance_dataset.groupby(["weekday","time_of_day"]).size().reset_index()) td_high_day.columns = ["Weekday", "Time of Day", "Count"]

plt.figure(figsize=(15, 10)) sns.lineplot(x="Time of Day", y="Count", data=td_high_day).set_title("Time of Day wise Distance");

td_high_day = pd.DataFrame(high_distance_dataset.groupby(["weekday","name","time_of_day"]).size().reset_index()) td_high_day.columns = ["Weekday", "name","Time of Day", "Count"]

plt.figure(figsize=(15, 10)) sns.lineplot(x="name", y="Count",hue="Time of Day",data=td_high_day,estimator=np.max).set_title("Time of Day wise Surge");