This "Hack the Feed" hackathon is a showdown to decode a treasure trove of social media data and create a comprehensive and reproducible report detailing my findings.
Participants are expected to: Create a comprehensive and reproducible report detailing their findings. Propose actionable recommendations based on the insights. Create a simple and engaging visualisation of your results & analysis.
Submissions will be evaluated based on the following:
Innovativeness: Originality and novelty of the insights.
Actionability: Practicality and feasibility of the recommendations.
Presentation Quality: Clarity and effectiveness in conveying findings in writing and visual form.
Depth of Analysis: How thoroughly the data has been explored and understood.
Begin by understanding the datasets provided. Explore the columns, types of data, and any missing values. Familiarize yourself with the platforms (Instagram, Facebook, Twitter, LinkedIn) from which the data originates to understand the context.
Handle missing values, either by imputing or removing them based on their relevance.
Convert data types if necessary (e.g., dates, categorical data). Remove any duplicate entries.
Use visualization tools to understand the distribution and trends in the data. Identify key metrics like engagement rates, post reach, and user demographics. Check for any patterns or anomalies in posting times, content types, or engagement.
What are the most engaging types of posts? Which platform yields the highest engagement for the client? What are the peak times for user engagement? Are there any noticeable trends over time (e.g., increasing likes, decreasing shares)?
Based on EDA, create new metrics or indicators that might be more informative. For instance, engagement rate as a ratio of engagements to impressions.
Use statistical methods or machine learning (if applicable) to understand underlying patterns or to predict future trends. Segment data to find insights specific to certain demographics or post types.
Based on the analysis, derive actionable insights. Translate these insights into recommendations for the client.
Use graphs, charts, and other visualisation tools to make your findings easily digestible. Ensure that visualisations are clear and can be understood by someone without a technical background. Provide a shareable link to the visualisation
At the end of the exercise, I am able to create a transformed data out of the raw data. I created a comprehensive and reproducible report detailing my findings. I created a simple and engaging visualization of my results & analysis.