Background
This project stems from the Accenture North America internship program on Forage. It revolves around a case study of a company called Social Buzz, which is a social media platform founded by two former engineers from a prominent social media company. Social Buzz prioritizes content by maintaining user anonymity and tracking reactions to content. With a user base of over 500 million monthly active users, they require assistance in managing their rapid growth and conducting data analysis. Their plan includes completing an Initial Public Offering (IPO) and seeking guidance, resources, and expertise from an advisory firm. The initial three-month project encompasses various tasks, such as auditing their big data practices, recommending IPO strategies, analyzing content categories, and providing best practices. The assigned responsibilities involve creating presentations, extracting data sets, auditing data centers, merging tables, conducting virtual sessions, preparing IPO best practice documents, loading data sets, conducting technology workshops and stress tests, communicating with previous IPO companies, and documenting the entire process.
The primary responsibility of the Data Analyst is to analyze the provided data sets and present visualizations to gain insights into the popularity of different content categories, particularly the top five.
To answer the business question and understand the interdependencies between tables, a data model was provided.
Data Cleaning and Modeling
The project involved three datasets: Reaction types, Reactions, and Content.
The Reaction types dataset included information about the Reaction Type, Sentiment, and Score.
The Reactions dataset included details such as Content ID, User ID, Reaction Type, and Date posted.
The Content dataset contained information about Content ID, User ID, Content Type (photo, video, audio or gif), Post Category, Post url
The data cleaning process encompassed the following steps:
Removal of unnecessary data, such as URLs, that do not contribute to answering the business problem.
Standardization of the 'Post Category' field, as it contained duplicate entries with some having quotation marks and others not.
Elimination of rows with missing values that would not adversely affect the project's objectives.
Standardization of the 'Date posted' column, as it was recorded in different formats, with some entries as text and others as dates.
The provided data model was replicated using Power Pivot, as illustrated below.
Additionally, a screenshot of the produced dashboard is provided below, and an interactive version can be downloaded from here. Please note that the downloaded file needs to be enabled for content.
Insights
The dataset comprises 16 unique post categories, including Food, Science, and Animals. 'Animals', 'Science', and 'Healthy eating' emerged as the most popular topics based on the number of posts and popularity scores. This suggests that the user base of Social Buzz is predominantly interested in Animals, factual content, and issues related to healthy eating/food. The significance of food interest is further reinforced by the inclusion of 'Food' among the next two popular categories.
Furthermore, the month of May witnessed the highest number of posts compared to other months throughout the year.
Finally, an engagement improvement campaign involving healthy eating brands may yield positive results.
Comments