Description
Analyzing Digital Marketing datasets
Goals
• To work with datasets using xsv(xcsv) &
Trifacta(https://community.trifacta.com/s/lesson-1-introduction-totrifacta-wrangler),
• Stage datasets in Snowflake
• To be able to analyze marketing data using Salesforce Einstein analytics
studio
• Derive insights from the datasets
• Crisply communicate and document your findings
Case
Marketa analytics has hired you to as an Algorithmic marketing analyst.
Marketa is a consulting organization specializing in Marketing analytical
solutions. Your client (see allocations by team number below) has provided
you a sample dataset and asked you to analyze and build an analytical
dashboard as a Proof-of-concept to illustrate the value of data driven
analytics. The themes to be considered could include:
• Pricing
• Promotion
• Search
2
• Recommendations etc.
Marketa wants you to analyze the data using tools (xsv, Trifacta, Snowflake)
and build a dashboard using Einstein analytics. They also want you to build a
codelabs document to crisply illustrate the value analytical solutions would
bring to the company. You are also asked to discuss what additional datasets
and methodologies could be used. The company has a challenge using large
scale datasets and are considering using Trifacta and xsv as data tools to
work with. You are expected to illustrate how you would:
• Use the tools for joining datasets
• Filtering
• Aggregating
• Missing value handling
• Deriving additional columns from existing datasets
• Cleaning (for example removing blank spaces, formatting dates,
Capitalizing etc.)
In order to do that you are asked to illustrate the strengths and weakness of
each tool/package
Dashboards:
Once you clean the data, import the data into Snowflake and illustrate how
to use the Einstein analytical dashboard to illustrate various aspects of
analysis. (https://salesforce-trailblazer.com/snowflake-einstein-analytics/).
Questions to consider:
• Which columns are dimensions, which columns are measures?
• How would you generate new dimensions? What will you do to
summarize measures?
• Who would use this dashboard?
• What value would generated using this dashboard ?
3
Deliverables:
• How to work with the large datasets using xsv and Trifacta
• Schemas for working with Snowflake for your chosen dataset
• Analytics Dashboard using Salesforce Einstein Analytics
• A Google Codelabs document summarizing the insights
Team allocations: (See the google sheet)
Instacart – Market Basket Analysis
The dataset is anonymized and contains a sample of over 3 million grocery orders from more
than 200,000 Instacart users.
• Data: https://www.instacart.com/datasets/grocery-shopping2017 or https://www.kaggle.com/c/instacart-market-basket-analysis/data
• Description: https://gist.github.com/jeremystan/c3b39d947d9b88b3ccff3147dbcf6c6b
• Backup copy (data and description): https://drive.google.com/drive/folders/1JCD3vtYI6iOSGaZ9DoSXDQ4GrvMzqLL
Criteo – Attribution Modeling for Bidding
This dataset represents a sample of 30 days of Criteo live traffic data. Each line corresponds to
one impression (a banner) that was displayed to a user. For each banner we have detailed
information about the context, if it was clicked, if it led to a conversion and if it led to a
conversion that was attributed to Criteo or not. Data has been sub-sampled and anonymized so
as not to disclose proprietary elements.
• Data: https://s3-eu-west-1.amazonaws.com/attributiondataset/criteo_attribution_dataset.zip
• Description: http://ailab.criteo.com/criteo-attribution-modeling-bidding-dataset/
4
• Backup copy (data and
description): https://drive.google.com/open?id=1WY6DdbbL6nzcxLA3z3vWYAbqNXeCg
9Qu
Dunnhumby – The Complete Journey
Household level transactions over two years from a group of 2,500 households who are
frequent shoppers at a retailer All of a household’s purchases within the store, not just those
from a limited number of categories Demographics and direct marketing contact history for
select households
• Data: https://www.dunnhumby.com/careers/engineering/sourcefiles
• Description: https://www.dunnhumby.com/careers/engineering/sourcefiles
• Backup copy (data and
description): https://drive.google.com/drive/folders/1PAe62y3fgxPSgzvkMph3295Ah9
WCMrhR
Yoochoose – RecSys Challenge 2015
The data represents six months of activities of a big e-commerce businesses in Europe selling all
kinds of stuff such as garden tools, toys, clothes, electronics and much more.
• Data: https://recsys.yoochoose.net/challenge.html
• Description: https://recsys.yoochoose.net/challenge.html
• Backup copy (data and
description): https://drive.google.com/drive/folders/1pQXY_Pl6UaLYcfvN92pqbDyk2auv
yibA
Kaggle – Give Me Some Credit
Historical data are provided on 250,000 borrowers and the prize pool is $5,000 ($3,000 for first,
$1,500 for second and $500 for third).
• Data: https://www.kaggle.com/c/GiveMeSomeCredit/data
5
• Description: https://www.kaggle.com/c/GiveMeSomeCredit/data
• Backup copy (data and
description): https://drive.google.com/drive/folders/14Ss_wSOHP8L7KmHxZelttTR6Oad
A2ELU
MovieLens – 25M Dataset
MovieLens 25M movie ratings. Stable benchmark dataset. 25 million ratings and one million tag
applications applied to 62,000 movies by 162,000 users. Includes tag genome data with 15
million relevance scores across 1,129 tags. Released 12/2019
• Data: https://grouplens.org/datasets/movielens/25m/
• Description: https://grouplens.org/datasets/movielens/25m/
• Backup copy (data and
description): https://drive.google.com/drive/folders/1GhJGkFAwNb95Jnah6OEKJH2oDg
0ls25g
Elo – Merchant Category Recommendation
This dataset is created by Elo, one of the largest payment brands in Brazil. The datset contain
contains up to 3 months’ worth of transactions for every card.
• Data: https://www.kaggle.com/c/elo-merchant-category-recommendation/data
• Description: https://www.kaggle.com/c/elo-merchant-categoryrecommendation/overview
• Backup copy (data and
description): https://drive.google.com/drive/folders/1HmrVX4nAT3AVD9jHIe_zpTKn7Jh-J-h?usp=sharing
Reference: https://github.com/ikatsov/tensor-house/blob/master/resources/datasets.md
