Scribd
Upload
36 views

E-Commerce-Data-Analysis

Uploaded by

Mridul
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
36 views

E-Commerce-Data-Analysis

Uploaded by

Mridul
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

# E-Commerce Data Analysis and Visualizations

An E-commerce company ran marketing campaigns to generate sales. They collected


information on the type of device each customer used, the sale amount, total transaction for
each day, etc. Here, we need to analyse the performance of the various marketing
campaigns and understood its relationship with respect to user interaction and total sales.
Finally, steps to improve campaign performance are discussed.

## Exploratory Data Analysis


### Data Dictionary
Data Dictionary explains what each column of the dataset represents

Column Name | Description


------------|------------
device_created_on | Device creation Timestamp
device_type | Type of Device
operating_system | Operating System on which user accessed the app
attribution_created_on | Attribution Campaign Timestamp
campaign | Marketing campaign name
user_id | User Id
user_created_on | User creation Timestamp
name | User name
amount | Transaction amount
sale_created_on | Sale creation Timestamp
sale_id | Sale Id
date | Sale created day
weekday | Day on which sale happened
item_id| Item Id
Task needs to be achieved:

1. Acquire the data, dump the data into some of the databases (SQL, Mongo DB,
Casandra local or cloud version)
2. Connect with the business user and try to get the understanding of the data
attribute
3. Connect with the business user and try to get the understanding about the KPI
(Key performance indicator)
A Key Performance Indicator (KPI) is a measurable value that demonstrates
how effectively a company is achieving key business objectives
4. Connect with the business user with raw visualization and gather user
experience and expectations feedback based on ease of use
5. Decide total number of dashboards based on user hierarchy and organization
6. Start building production-based dashboard
7. Below are the KPI which need to be captured

1. Relationship between Device Type and User Creation


2. Relation between Marketing Campaigns and User creation
3. Relation between Campaigns and Weekday
4. trend between sum amount of sale for top 10 campaigns across weekdays.

1. Send the Dashboard for a review for the stockholder


2. Performed UAT (user acceptance testing)
3. Go for the random test
4. Make it live
5. Share link and authorization for the user
6. Keep it in Hypercare for any modification

Lastly, as a chronic over-achiever:

 Find at least two unexpected phenomena in the data and provide a


visualization and analysis to document their presence.

Considerations
Remember, the people reading your analysis will NOT be data analysts. Your
audience will be general public . Your data and analysis need to be presented in a
way that is focused, concise, easy-to-understand, and visually compelling. Your
visualizations should be colourful enough to be included in press releases, and your
analysis should be thoughtful enough for dictating programmatic changes.

Assessment
Your final product will be assessed on the following metrics:

 Analytic Rigor

 Readability

 Visual Attraction

Hints
 You may need to get creative in how you combine each of the CSV files. Don't
just assume Tableau is the right tool for the job. At this point, you have a
wealth of technical skills and research abilities. Dig for an approach that works
and just go with it.

 Don't just assume the CSV format hasn't changed since 2013. Subtle
changes to the formats in any of your columns can blockade your analysis.
Ensure your data is consistent and clean throughout your analysis. (Hint: Start
and End Time change at some point in the history logs).

 Consider building your dashboards with small extracts of the data (i.e. single
files) before attempting to import the whole thing. What you will find is that
importing all 20+ million records of data will create performance issues
quickly. Welcome to "Big Data."

 While utilizing all of the data may seem like a nice power play, consider the
time-course in making your analysis.

 Remember, data alone doesn't "answer" anything. You will need to


accompany your data visualizations with clear and directed answers and
analysis.

 As is often the case, your clients are asking for a LOT of answers. Be
considerate about their need-to-know and the importance of not "cramming in
everything". Of course, answer each question, but do so in a way that is
organized and presentable.

 Keep a close eye for obvious outliers or false data. Not everyone who signs
up for the program is answering honestly.

 In answering the question of "why" a phenomenon is occurring, consider


adding other pieces of information on socioeconomic or other geographic
data. Tableau has a map "layer" feature that you may find handy.

 Don't be afraid to manipulate your data and play with settings in Tableau.
Tableau is meant to be explored. We haven't covered all that you need -- so
you will need to keep an eye out for new tricks.

 The final "format" of your deliverable is up to you. It can be an embedded


Tableau dashboard, a Tableau Story, a Tableau visualization + PDF -- you
name it. The bottom line is: This is your story to tell. Use the medium you
deem most effective. (But you should definitely be using Tableau in some
way!)

 Treat this as a serious endeavour! This is an opportunity to show future


employers that you have what it takes to be a top-notch analyst.

 Good luck!

REQUIREMENTS
Submissions must meet the following requirements:
 Include a Project built with the required developer tools and meets the
above Project Requirements.
 Include a text description that should explain the problem your Project
is attempting to solve and its features and functionality.
 Include a demonstration video of your Project. The video portion of the
submission:

 should be less than three (3) minutes


 should include footage that shows the Project functioning on the device
for which it was built
 must be uploaded to and made publicly visible on YouTube and a link
to the video must be provided.
 must not include third party trademarks, or copyrighted music or other
material unless the Entrant has permission to use such material.

 Include a URL to a code repository on GitHub or another code


repository platform. If the code repository is private, Entrant must
provide access to the GitHub account
 Include a list of the APIs and Development tools used within the
project.
 Include potential further improvements to your Project if more time
were permitted.
 Be the original work of the submitter, be solely owned by the submitter,
and not violate the IP rights of any other person or entity.
 Multiple Submissions: An Entrant may submit more than one
Submission; however, each submission must be unique and
substantially different from each of the Entrant’s other Submissions.

You might also like