I had the privilege of participating in the open-access Accenture Data Analytics Virtual Experience Program, offered through Forage. During this engaging experience, I assumed the role of a data analyst, collaborating with ‘Social Buzz,’ a dynamic organization. My primary task was to delve into their extensive data repository and provide insights to empower them in harnessing the full potential of their vast data resources
Project Overview and Business Problem:
Client’s Name: Social Buzz
Client’s Industry: Social media & Content creation
Client’s Background: Social Buzz, a rapidly expanding social media organization, has garnered an impressive user base of over 500 million active users per month in just 5 years. They’ve outpaced their growth projections and are seeking the guidance of an advisory firm to efficiently manage their scaling process.
In light of their swift expansion and the digital nature of their core product, Social Buzz grapples with substantial data generation, collection, and the need for in-depth analysis. Each day, they receive more than 100,000 pieces of content, encompassing text, images, videos, and GIFs. This data is predominantly unstructured and demands sophisticated and costly technology for proper management and upkeep.
Historically, they’ve relied solely on their in-house resources to achieve their current standing. However, one of the primary motives for considering external expertise is their desire to adopt data best practices employed by industry leaders. Given the immense volume of data they handle, they’re eager to gain insights into how the world’s largest corporations tackle the challenges of big data.
To kick off our engagement with Social Buzz, we are embarking on a three-month initial project to demonstrate our prowess. Social Buzz’s expectations for this project include:
- An audit of their big data practices
- An analysis of their content categories that highlights the top 5 categories with the largest aggregate popularity.
Business problem: The client has reached a massive scale within recent years and does not have the resources internally to handle it.
About Dataset:
Content Dataset: The Content Dataset comprises 1,000 rows and 5 columns. It provides details about each post, including the content type, which can be photos, videos, or other media forms. Additionally, the dataset classifies posts into various categories such as animals, food, science, and more.
Reactions Dataset: The Reaction Dataset consists of 25,553 rows and 4 columns. It primarily focuses on recording the types of reactions elicited by posts and their corresponding date and time of occurrence.
ReactionType Dataset:The Reaction Types Dataset provides information on various reaction types received by posts. It includes data on the type of reaction, associated scores for the posts, and sentiment types.
Note: If you wish to conduct your own analysis, you can access the dataset here.
Content , Reactions, ReactionTypes
The project utilized the following tools:
MySQL Workbench: This software played a pivotal role in data management, encompassing data cleansing and modeling. It served as the primary platform for data processing and analysis.
Microsoft Excel: Microsoft Excel played a pivotal role in data analysis, aiding in the extraction of key insights from the datasets and creating charts for visualization.
Microsoft PowerPoint: Microsoft PowerPoint was employed to craft presentation slides for client meetings, ensuring clear and organized communication of findings and recommendations.
Throughout this experience, I have applied my data analytics skills to:
Data Cleaning and Modeling: I proficiently cleaned and conducted data modeling on the client’s datasets using MySQL Workbench, ensuring data accuracy and integrity.
Data Analysis: I carried out a comprehensive analysis of the client’s data to identify the top 5 content categories with the highest aggregate popularity.
Data Visualization: Leveraging Microsoft Excel, I created engaging data visualizations that effectively represented key insights derived from the analysis.
Storytelling with Data: I adeptly conveyed the narrative within the data by using the visualizations to present meaningful insights.
Presentation: I prepared a compelling presentation for the client using Microsoft PowerPoint, highlighting and sharing valuable insights extracted from the data analysis.
Data Cleaning and Data Modeling :
Throughout the data cleaning process, I meticulously attended to various aspects of the datasets, performing all operations in MySQL Workbench.
Data cleaning was done separately, starting with the removal of blank entries to ensure data integrity. Subsequently, spelling errors were corrected to enhance data quality and maintain consistency. Unnecessary columns were removed to keep only the most important data. I also changed the column names to make the data easier to understand.
To create a comprehensive dataset for analysis, I aggregated data from three distinct datasets using SQL JOIN operations within MySQL Workbench. This consolidation process resulted in a unified dataset comprising 24,573 rows and 8 columns. This consolidated dataset served as the foundation for subsequent data analysis and visualization, allowing for the extraction of valuable insights
You can explore my SQL Workbench code by visiting my GitHub Repository here. DataCleaning , DataModelling
You can access the final dataset for analysis here
Data Analysis and Data Visualization:
In our data analysis, we discovered the top 5 most popular content categories on Social Buzz. ‘Animals’ leads the chart with a score of 74,965, followed by ‘Science,’ ‘Healthy Eating,’ ‘Technology,’ and ‘Food.’ This insight guides our content strategy to align with these categories, creating more engaging content. ‘Animals’ also garners the highest engagement, emphasizing its appeal. ‘Science’ follows closely, highlighting users’ intellectual curiosity
Our percentage share chart on the left reveals content engagement on Social Buzz. ‘Animals’ leads with 21.36%, followed by ‘Science’ at 20.28%, ‘Healthy Eating’ at 19.76%, ‘Technology’ at 19.59%, and ‘Food’ at 19.00%. This guides our content strategy to meet user preferences.
The right-side pie chart displays content engagement by type. Photos lead with 27%, followed by GIFs at 24.47%, videos at 25.31%, and audio at 23.22%. It demonstrates balanced engagement across content types, showcasing diverse user interests.
The below line chart tracks post trends by month. January leads with 2,126 posts, reflecting the festive spirit and extended holidays. In contrast, February has the lowest count at 1,914 posts due to its shorter duration and post-holiday slowdown. This highlights the influence of seasons and celebrations on user sentiment.
Storytelling with Presentations:
In our presentation, we’ll guide you through a compelling data-driven story:
- Delve into the diverse engagement by content type, revealing user preferences in photos, GIFs, videos, and audio.
- Understand how the season impacts user sentiment and post trends, with holidays and festivities influencing content.
- Ultimately, these insights equip us to tailor content, enhancing user engagement and satisfaction.
For the complete presentation and a deeper understanding of our findings, please access the presentation Here. We’re excited to refine Social Buzz’s content strategy based on these valuable insights.
Recommendations:
- Capitalize on the popularity of ‘Animals’ and ‘Science’ by consistently creating “real-life” and “factual” content in these categories to maintain and enhance user engagement.
- ‘Healthy Eating’ is a prominent theme within the top 5 categories. Consider launching a targeted campaign and collaborating with healthy eating brands to further boost user engagement, leveraging this audience interest.
- The presence of ‘Technology’ content indicates user enjoyment in this category. Explore partnerships with tech industry leaders to significantly increase user engagement, aligning with the rise in technology trends.
Thank you to all readers for your time and interest in this project.