Enhancing Social Buzz Scaling Process: A Comprehensive Data Analysis Approach
INTRODUCTION: Accenture Data Analytics and Visualization Virtual Internships
In the realm of my recent Virtual Internship Program, an exciting opportunity emerged as a client reached out to the advisory firm I had the privilege of interning at — Accenture. @Forage. The client, Social Buzz, a prominent player in the domain of social media and content creation, sought our expertise to analyze their extensive dataset and ensure a seamless scaling process. As a data analyst, I undertook the responsibility of this intricate task, employing diverse methodologies to extract precise insights.
CLIENT OVERVIEW:
Our client, Social Buzz, has made remarkable strides within the social media and content creation sector, charting an impressive growth trajectory over the past half-decade. With an astounding monthly active user base exceeding 500 million, their rapid ascent necessitated adept guidance to manage their scaling endeavors. Thus, our collaboration with Social Buzz, marked by our engagement with their data intricacies, was initiated.
DATA LANDSCAPE AND CHALLENGE:
In the dynamic landscape of Social Buzz, characterized by its digital core, an immense volume of data is generated, collected, and subsequently analyzed. The daily influx of over 100,000 content pieces, spanning text, images, videos, and GIFs, underscores the magnitude of their data challenge. Notably, this data assumes an unstructured nature, demanding sophisticated and resource-intensive technologies for effective handling and maintenance.
PROJECTS SCOPE AND OBJECTIVES:
Our partnership commences with a focused three-month project, emblematic of our commitment to proving our mettle to Social Buzz. The project’s scope encompasses pivotal areas:
1.Big Data Audit: A meticulous evaluation of Social Buzz’s big data practices, serving as a foundation for actionable insights.
2. Strategic IPO Recommendations: Offering expert recommendations to bolster the potential for a successful Initial Public Offering (IPO).
3. Content Category Analysis: A comprehensive analysis of content categories, spotlighting the top five categories that wield the highest aggregate popularity.
TASK ASSIGNMENTS:
My mission is facilitated through a well-structured set of tasks, each designed to address specific facets of the project:
1.Best Practices Presentation: Crafting an up-to-date presentation on prevailing big data best practices to guide Social Buzz’s endeavors.
2. Data Extraction and Integration: Employing SQL, Excel, or Power BI to extract and harmoniously integrate sample datasets.
3. On-site Data-Center Audit: Conducting an in-depth audit of Social Buzz’s data-center infrastructure to ensure operational efficiency.
4. Data Set Consolidation: Seamlessly merging distinct sample dataset tables to provide a holistic view of the data landscape.
5. Client Engagement Session: Facilitating a virtual engagement session, wherein we present past client success stories pertinent to Social Buzz’s aspirations.
So how do I approach this problem?
well I solved it in 5 steps:
1. Data preparation and understanding — The key to success on any data project is to understand the data in detail. So I took the time to understand the data model and domain of the business.
2. Data cleaning — After understanding the business, I then cleaned the available datasets and thought about what an ideal dataset should look like for this problem.
3. Data modelling — After ensuring the data was clean for analysis, I needed to process and model this data into a dataset that can precisely answer the business questions and produce the results needed.
4. Data analysis — With the new dataset, I used my analytical expertise to uncover insights from this dataset and to produce visualizations to describe the insights.
5. Recommendations — And finally I used these insights to unlock business decisions and to make recommendations on next steps.
Tools Used In carrying out this project: Microsoft Excel, Microsoft Power Point and Power BI
Collection Requirements:
I was presented with seven (7) datasets. Among them, only three data sets are relatively necessary to answer the commercial questions.
- Data Preparation and Understanding:
During this process, I understood the client’s need and used data modeling techniques to select the three most needed datasets. Here are the techniques given. Three datasets were selected as contents, reactions and reaction type. Because we need data including content id, category, content type, reaction type, sentiments, date-time and reaction score.
2. Data cleaning:
The second step in our analysis was data cleaning, where I diligently removed null and blank values, standardized column names, and excluded irrelevant data columns, streamlining the datasets for further analysis.
3. Data Modeling:
Next, I seamlessly integrated the Contents and Reactions Type datasets with the Reaction table using the VLOOKUP formula, allowing for a comprehensive view of the relevant information.
- To gain valuable insights, I created a new sheet titled “Aggregate Categories.” In this process, I extracted unique categories by eliminating duplicates from the Categories column. Utilizing the SumIf formula, I calculated the total sum of reactions for each category and ranked them based on performance, revealing the top 5 performing categories.
4. Data Analysis Insights And Visualizations:
( A ) There were 16 distinct categories, 4 content types, and a total emotion score of 974 000.
( B ). Positive (Green), Neutral (Yellow), and Negative (Red) are the three colors that represent the feelings score. There have been 824K favorable reactions, 85K neutral reactions, and 65K negative reactions to the materials that have been submitted.
( C ). The leading content categories, based on popularity percentages, encompass a diverse range, including animals 21.36%, science 20.28%, healthy eating 19.76%, technology 19.59%, and food 19%. Each category boasts a significant share of the content landscape, contributing distinctively to the overall engagement.
( D ). In terms of content distribution across months, May emerges as the frontrunner with the highest posting frequency, totaling 2136 posts. In contrast, February witnessed a slightly lower count, tallying 1917 posts. This divergence underscores the notable disparity in viewership, indicating a heightened interest in the content shared during the month of May compared to February. These transient fluctuations mark pivotal shifts in the patterns of viewership engagement.
( E ). This analysis reveals the diverse modalities through which content was shared, encompassing photos, videos, GIFs, and Audio. Among these, the photo content type takes the lead, boasting the highest posting frequency at an impressive 6.6k content items. Close on its heels is the video content type, contributing substantially with 6.2k posts. GIFs occupy the third position with 6.1k posts, while audio content type brings up the rear, yet still significant, with a total of 5.7k posted items.
( F ). Within this section, the aggregation of scores across the 16 distinct categories comes into focus. Upon closer scrutiny of the reaction categories, “Animals” and “Science” emerge as the predominant leaders, accumulating 75,000 and 71,000 viewed posts, respectively. Following closely are “Veganism” and “Public Speaking,” amassing a combined total of 50,000 and 49,000 views, respectively.
( G ). Dashboard: This is a Pictorial representation of dashboard and its visualizations.
5. Recommendations:
- Shifting my focus to the crux of the analysis, the standout categories that command the most attention are unequivocally “Animals” and “Science.” Consequently, I highly recommend channeling your efforts towards cultivating content centered around these genres. By meticulously crafting captivating content and amplifying promotional endeavors, you stand to usher in a fresh influx of viewers, while simultaneously retaining your existing audience. This concerted approach promises to augur heightened engagement levels and an enriched entertainment experience.
- On an alternative note, it’s noteworthy that “Food” content seemingly struggles to command a comparable viewership percentage within the upper echelons of the top 5 categories. To surmount this challenge, I suggest fostering synergistic partnerships with food brands, possibly complemented by the production of live-streamed culinary content. This dual-pronged strategy holds the potential to resonate with your audience, effectively rekindling interest and bolstering engagement within the “Food” category.
- Additionally, a similar phenomenon is observed with “Public Speaking” content, where viewer engagement appears less pronounced. To mitigate any waning enthusiasm, consider interspersing memes between posts. This prudent tactic has the potential to sustain viewer captivation and mitigate the risk of viewer attrition, which can sometimes be inherent to content of this nature. Furthermore, when crafting video content, a succinct and compelling approach is advised, as protracted content may inadvertently lead to viewer disengagement.
CONCLUSION:
The collaboration with Social Buzz, facilitated by Accenture’s Data Analytics and Visualization Virtual Internships, has unveiled a wealth of insights to propel their growth. A meticulous five-step methodology allowed me to dissect the data landscape, revealing the dominance of “Animals” and “Science” content categories. Additionally, I’ve deciphered viewer sentiments, identified peak engagement months, and highlighted content type dynamics.
This analysis has not only enriched my understanding of audience behavior but also laid the foundation for strategic recommendations. By spotlighting key categories and offering tailored suggestions, I am poised to guide Social Buzz toward data-driven success. This journey underscores the transformative potential of data analytics, solidifying our commitment to steering Social Buzz’s trajectory toward sustained growth in the dynamic landscape of social media and content creation.
#Data Analysis #Data Visualizations #Virtual Internships #Excel #Power BI
