Table of Contents
Social Buzz Case Study
Introduction
Currently nearing the completion of my Google Data Analytics certificate, I have selected the Social Buzz data set for a comprehensive case study. The main purpose is to exhibit my abilities in data cleaning and analysis, showcasing my mastery of the "Ask, Prepare, Process, Analyze, Share, and Act" methodology.
Ask
About Social Buzz
Social Buzz is a content-centric social media platform founded by former engineers from a major social media conglomerate. Emphasizing anonymity and diverse user reactions, the platform prioritizes trending content over individual identities. With over 500 million monthly active users, Social Buzz has experienced remarkable growth, necessitating expert guidance to manage their scaling process effectively. Due to their rapid growth, they are asking Accenture to audit their big data practice, give recommendations for a successful IPO, and analyze their content categories– they are interested in the analysis of their content categories, but to also find out what their most “popular” ones are based on a qualitative scoring system.
Questions for Analysis
What are the top 5 content categories based on “popularity” score?
What opportunities arise to manage the data from the initial study?
What trends and opportunities can Social Buzz capitalize on?
Business Task
The task focuses on analyzing the most popular content categories on the social media platform to identify optimal data practices for handling the substantial daily data influx. This analysis allows for the recognition of opportunities to leverage.
Prepare
Data Set
The data source for our case study is Accenture. The dataset is stored on Forage to complete an analysis, presentation, and with the final step to present the findings.
Data Credibility
It’s a small sample size of over 25,000 reactions, labeled by 16 qualitative reaction types. The data set focuses on 16 content category types. Social Buzz’s platform has over 100 categories of content, which might allow for better analysis if there are overlapping and duplicate category types once the entire data structure is reviewed.
Importing Data Sets
For this case study, I will be using Excel to aggregate the data. So I import the CSV files into a workbook to review and begin.
Data Sets
The data sets provided are within 3 CSV files, but for my analysis I will need to aggregate them into 1 with my insights on the data on a separate sheet. The originating CSV files were Reactions, Content, and ReactionTypes, which I imported into Excel to be able to aggregate them properly. I renamed my final aggregate data to “Aggregate Data” and had a separate “Insights” tab for my final findings.
Process
Cleaning Data Sets
Reactions:
Added column filters
Removed User ID column
Removed incomplete data– deleted any rows with blanks in Type column using the filter
Extracted timestamps in the DateTime column using Text to Columns to keep date only and reformatted, then changed header to only “Date” to avoid any confusion
Content:
Added column filters to audit data easily, then removed User ID and URL columns
Unified Category results by removing quotation marks
Made data uniform by using lower case
Changed dogs to animals to create cohesion between the 2 categories
ReactionTypes:
Fixed/ swapped neutral and positive type to keep score progression
Fixed duplicate reaction scores – prioritized “want” over “cherish” for possible advertising opportunities
Aggregation
Added 4 columns to Reactions, created more unique column attributes for clarity
Used Vlookups to pull in applicable data
Content: Type and Category based on Reactions’ Content ID
ReactionTypes: Sentiment and Score based on Reactions’ Type
Hardcoded formulas to allow for further manipulation of data and removed Content and ReactionTypes sheets
Improved navigation
Froze top row to keep column headers on top
Renamed sheet to Aggregate Data
Increased legibility by formatting column headers
Reordered columns to establish a more natural flow of topic
Analyze
First Insights
While this was a small data set, there does seem to be an opportunity to streamline the content categories– for example, removing “dogs” as its own category and rolling those in under “animals”. While there are similar topics like cooking, healthy eating, food, and veganism are related, they potentially could have vastly different target demographics. For instance, veganism does incorporate food but it also encompasses lifestyle.
Focus
Streamline content categories by consolidating categories like "dogs" under the broader category of "animals" to improve categorization efficiency.
Recognize the potential variations in target demographics among similar topics like cooking, healthy eating, food, and veganism. Consider refining these categories to better capture their distinct audiences.
Leverage the uniqueness of Social Buzz's over 100 reaction types by mirroring the two-tier categorization process in content topics. This approach reduces the number of metrics to analyze when examining the company's overall data.
Focus on qualitative data by prioritizing positive reaction types over negative ones, using the popularity score as a guiding factor.
Share
Visual Data Telling
Created new sheet for pivot tables, focusing on primary question of what the top 5 categories are by popularity score
Changed workbook style to keep with presentation branding colors
Added pivot table charts to visualize data, including further insights explored which highlighted key takeaways from each section
Filtered applicable pivot tables for the top 5 categories by popularity
Improved navigation & visual queues
Highlighted top results for quick reference
Added top “headers”
Alternated colors of pivot tables to group “topics” together
Added collapsible handles to view or compare specific topics
Kept color pallet of charts limited to refrain from asking the viewer to interpret or learn complex color keys
Presentation
Edited Accenture presentation to keep in with brand themes and color palette, increased readability of all previously present text
Highlighted key information like key phrases or words to engage stakeholders
Created engaging, animated charts and slides to pique interest
Conclusion
Summary
Animals and science are among the top 5 categories based on popularity score, showing that there is a high engagement with “factual” and “real-life” content. Additionally, healthy eating, along with other food and related topics such as culture, cooking, and travel, score high in popularity. This suggests a strong interest in food and lifestyle content among Social Buzz’s user base.
Actionable Items
Capitalize on the strong interest in healthy eating and related topics like culture, cooking, and travel among Social Buzz's user base.
Forge partnerships with reputable healthy food brands to create engaging recipe and cooking content that features sponsored ingredients and products.
Emphasize authenticity and excitement in the content to resonate with users seeking food and lifestyle-related content.
Explore brand collaborations in the culinary-tech space, taking advantage of the intersection between technology and the top 5 categories.
Continuously monitor and analyze user reactions and feedback to refine content strategies and optimize user engagement.