You are on page 1of 9

Kateryna Miniailo ICA

Element 1
Data Science at Bolt
This report investigates the role of data science in enhancing business
performance through a case study transportation platform called Bolt.
To begin with, Bolt uses data science to predict demand for rides in different
locations and at various times. This helps them optimize pricing and allocate
drivers efficiently, reducing waiting times for passengers during peak hours and
ensuring adequate coverage during low-demand periods.
Bolt wants to improve user retention by understanding rider behaviour patterns
and identifying factors that influence riders to continue using their platform
over time.
Bolt collects various data points for each ride and rider, such as: ride
timestamps, pick-up and drop-off locations, distance and duration of rides,
ratings and feedback given by riders, rider demographics (age, gender, etc.),
frequency of app usage, ride history (number of rides taken in a given period),
location-based data (city, region, etc.).
Using demographic data and ride history, riders can be segmented into different
groups based on their frequency of app usage and ride patterns. This helps
identify loyal, occasional, and churned users.
Data scientists at Bolt analyse rider ratings and feedback to identify common
issues or pain points experienced by users during their rides. This provides
insights into areas where improvements are needed to enhance the user
experience. Then, they calculate the overall user retention rate and track it over
time. This helps understand if there are certain periods or events that impact
user retention positively or negatively. Next goes analysis of the number of
rides taken by individual riders over a specific time frame. This helps identify
patterns of increased or decreased ride frequency, which may be linked to
specific events or promotions.
Comparing user behaviour and retention rates across different cities or regions
can reveal geographical differences in rider preferences and help tailor services
to specific markets.
Data scientists identify which app features are most frequently used by riders.
This information can be used to optimize the app's user interface and prioritize
feature enhancements. They also investigate factors that contribute to rider
churn (users who stop using the app). This includes looking at ride patterns,
ratings, and feedback from riders who have churned to identify potential reasons
for their disengagement.
Based on the analysis of ride experience and user feedback, Bolt can focus on
improving specific aspects of the service that have been identified as pain
points. For example, addressing issues related to driver behaviour, route
optimization, or app functionality.
By understanding user segmentation and ride frequency patterns, Bolt can
create targeted promotions and offers to encourage occasional users to become
more frequent users and retain loyal customers.
Comparing user behaviour across different cities allows Bolt to implement
localized strategies to better meet the preferences and needs of riders in specific
regions.
The churn analysis helps Bolt to implement preventive measures to reduce user
attrition. For instance, they could offer personalized incentives to users who
show signs of disengagement or address recurring issues faced by churned
users.
In the past, Bolt used manual decision-making, had static pricing and handled
the feedback manually which limited Bolt's ability to make data-driven
decisions, understand user behaviour at a granular level, and optimize their
operations fully. The introduction of data science has significantly changed the
way businesses like Bolt operate by enabling them to harness the power of data
to make more informed and efficient decisions, enhance the user experience,
and drive business growth.
References:
Denis Kovalenko “Scaling Data Science at Bolt”:
https://www.youtube.com/watch?v=UhTlwCu7hZY&ab_channel=fwdays
Bolt Labs Blog on Medium: https://medium.com/bolt-labs

Element 2
1. Implementing programming techniques in processing the dataset:
Store a column in a vector:
Create a list and assign multiple columns to it:

Create a data frame consisting first 4 columns and 30 rows from the entire
dataset:
Find the attributes/properties (e.g. type, length, rnames, names etc.) for each
data structure above and identify the difference
2. Use the R (https://cran.r-project.org/doc/manuals/r-release/R-intro.html) or
https://www.w3schools.in/r/ to further practice on R programming. Show one
example for each using RSudio environment. (20 marks)
R Operators:
Arithmetic and relational:

Logical:

Assignment:
Miscellaneous:

Conditional Statements:
If:

Else:

Else if:

Functions:
Built-in:
User-defined:

Removing missing/NA values from data:

You might also like