Uber Trips

Exploratory Data Analysis

Intro

Uber has changed the way I move around the world. It has made hailing a taxi fast and comfortable. Also, It has promised to make travel more efficient by using machine learning powered routing, ride splitting, dynamic pricing and managing a good supply of drivers .

However, in light of studies that have shown that Uber has worsened traffic congestion, I would like to analyze my trip data to see if Uber is really a reliable mobility solution or just cannibalizing walk-able or bike-able trips.

Questions of interest

How much is my average trip?

Which trips are my most popular?

What is the average trip duration and distance?

What percentage of trips have a distance of 7.5km or less (bike-able) and 3km or less (walk-able)?

How much did I spend on bike-able trips and what would the payback period be had I just bought a bike and rode it?

What percentage of trips have a duration of less than 10 minutes?

What percentage of trips get cancelled?

How far could I walk/cycle during my trip wait time?

map1.add_data(data=df, name='tripdata') map1

Reliability

Uber is quite a reliable service, with a 1.78% driver cancellation rate and 4.99 rider cancellation rate.

Trip Distance Distribution

The distribution of my trip distances are not symmetrical but rather skewed to the right. Most trips are less than 10km, which indicates frequent travel within the same city.

Product Types

The distribution of my fare amounts are skewed to the right. Most of my Uber expenditure is less than 200 ZAR. Uber has included UberEATS Marketplace expenditure within the trips_data.csv file. Unsure whether this is just the delivery fee or the actual full fee including food.

Further analysis reveals that UberEATS Marketplace data includes the cost of the food ordered, thus it will be dropped.

Other than UberEATS orders, all other product types seem to have a strong positive correlation between distance and fare amount.

Trip Statistics

Average distance of trips (km): 6.01

Median distance of trips (km): 5.15

75th percentile distance of trips (km): 6.87

25th percentile distance of trips (km): 2.08

Total distance of trips (km): 1828.36

Total spent on trips in South Africa (ZAR): 17362.16

Price per km traveled (ZAR): 9.50

Mean cost per trip (ZAR): 57.11

Aside by a few outliers, most trips are between 1km to 7km.

Bike-able and Walk-able

Walk-able trips are 3km or less, being 34.21% of trips while bike-able trips are 7.5 km or less, which were 82.57% of my trips.

A majority of my trips are not walk-able but a significant chunk are (approx 150).

An overwhelming majority of my Uber trips are bike-able, at around 300 trips.

Wait Times and Trip Duration

The data-set does not have a trip duration variable, therefore we will create one using the following code.

Trip Speed

The average speed while inside the uber taxi is 30.76 km/h. If I include the wait time into the calculation, the average speed decreases significantly to 18.71 km/h.