This project analyzes a social media dataset to answer six questions assigned.
This repository contains all code, documentation, and visual outputs required for the assignment.
Data: team09_social_media.csv
Notebook: DataChallenge2P08.ipynb
README.md
- Python 3
- Pandas
- Numpy
- Google Collab
- Visual Studio Code
- Github
- Loaded Dataset into Pandas
- Reviewed structure and data types
- Identified missing values
- Checked for invalid or inconsistent entries
- Examined summary stats
- Handled mising values
- 74 missing likes, 50 missing follower_count
- We chose to fill them in with the median since social media engagement often contains outliers that we skew the mean.
- Converted incorrect data types
- All that we had to correct was the time/date format on post date.
1. Which post type receives the most engagement (likes + comments + shares)? Photo posts recieed the most engagement on average with an average of 124.13, the next runner up being the "link" at 121.37.
2. What time of day generates the highest average likes? Afternoon posts recieved the highest average likes with a average like count of 109.35, while the runner up having an average of 100.73 Morning
3. How does follower count correlate with engagement? There is almost no correlation between follower count and engagement since the correlation value is 0.0139.
4. What is the optimal number of hashtags for maximum engagement? 9 at first shows up as the best engagement, but due to it being an outlier with only 1 post we cannot accept that as an answer. When we limit our consideration to atleast 5 posts, there is a answer of '0' with 128.82 or '6' with 128.58.
5. Which post type has the highest share rate? Videos posts are the most shared with an average share of 4.7, next runner up being the text with 4.61.
6. How does engagement vary by month? July has the highest average engagement with an average of 142.10, the runner up being April with an average of 127.72