uber-exploratory-data-analysis

Context

Ridesharing provides a way to quickly arrange transportation. It operates in an unpredictable market where demand can change drastically based on factors like time, location, weather conditions, and local events. Success in this industry hinges on identifying trends within these changes and meeting customer needs as they arise.

Objective

Uber Technologies, Inc., an American multinational ride-hailing company headquartered in San Francisco, operates in more than 785 cities globally and serves over 110 million users. As a recently appointed (imaginary) Data Scientist at Uber's New York office, my role involves analyzing data to uncover insights that can drive business growth.

Key Questions

• What are the various factors that impact the frequency of ride pickups?
• Which element has the most significant effect on the number of pickups, and why might this be?
• What strategies would I suggest to Uber's management to take advantage of the changing demand levels?

Data Description

The dataset includes information on Uber rides in different boroughs (subdivisions) of New York City, broken down by hour, along with details about the weather conditions at those times.

• pickup_dt: Date and time of the pick-up
• borough: NYC's borough
• pickups: Number of pickups for the period (hourly)
• spd: Wind speed in miles/hour
• vsb: Visibility in miles to the nearest tenth
• temp: Temperature in Fahrenheit
• dewp: Dew point in Fahrenheit
• slp: Sea level pressure
• pcp01: 1-hour liquid precipitation
• pcp06: 6-hour liquid precipitation
• pcp24: 24-hour liquid precipitation
• sd: Snow depth in inches
• hday: Being a holiday (Y) or not (N)

Source: https://www.kaggle.com/datasets/fivethirtyeight/uber-pickups-in-new-york-city

Actionable Insights and Recommendations

Insights

We analyzed nearly 30K hourly Uber pickup records across New York boroughs, spanning every day of the first six months of 2015. The primary focus of this analysis was on the number of pickups. From both environmental and business perspectives, it's inefficient to have excess cars in areas with low demand or insufficient cars during peak hours. We aimed to identify the factors influencing pickups and the nature of their effects.

Key conclusions from our analysis:

1. Uber cabs are most popular in Manhattan.
2. Weather conditions surprisingly do not significantly impact the number of Uber pickups.
3. Uber demand has steadily increased from January to June.
4. Pickup rates are higher on weekends compared to weekdays.
5. New Yorkers rely on Uber for leisure outings in the evenings.
6. Uber is also a popular choice for morning and evening commutes, with demand peaking between 7-8 PM.
7. Monday shows unusually low demand for Uber, which warrants further investigation.

Recommendations for Business

1. Focus on Manhattan, which is the most mature market for Uber. However, Brooklyn, Queens, and the Bronx also show potential for growth.
2. Maintain the growth momentum observed in the past months.
3. Ensure adequate cab availability during peak commuting hours on weekdays and late evenings on Saturdays.
4. Prioritize cab availability on Saturday nights when demand is highest.
5. Collect data on fleet size to better understand the demand-supply balance and develop a machine learning model to predict hourly pickups more accurately.
6. Gather more data on pricing to build a model that predicts optimal pricing strategies.

Further Analysis Recommendations

1. Investigate variations in cab demand between working days (weekdays) and non-working days (weekends and holidays).
2. Consider dropping boroughs with negligible pickups from the analysis to focus on areas with significant activity and uncover deeper insights.