ANLY482 Team wiki: 2015T2 TeamROLL Project Overview
Description | Methodology | Technology |
About SGAG
Project Motivation
Although SGAG has been very successful thus far, it also recognises that the online content space is very competitive, with newer players such as "SMRT Feedback" joining in the fray to generate local humour content. As such, SGAG needs to evaluate and improve their content strategy to ensure they stay entertaining and engaging to their audience.
Business Questions
In a nutshell, SGAG's business questions are:
- How can we drive audience growth through our creative content?
- What are the characteristics of a “great” post which appeals to our audience?
SGAG has so far thrived on an intuitive understanding of their customer's content preferences. However, they do not have a concrete or clear picture of the kinds of attributes which they can work on to make a specific post a "great" one. They hope that this project will utilise their rich pool of historical data to derive actionable insights into the concerns above, and offer recommendations how SGAG can improve their content creation strategy.
Analytical Objective
Taking into account the analytical problems faced, we aim to:
(i) Identify better performing posts and weak performing posts
- Based on clustering engagement level (number of likes, shares and comments)
(ii) Identify the patterns between the characteristics of a picture post and engagement level
- Based on the relationship between engagement level and picture design
- Picture design is based on the number of description lines, character used (e.g. Animals, Foreign celebrities) and the number of frames
- Picture design is based on the number of description lines, character used (e.g. Animals, Foreign celebrities) and the number of frames
(iii) Identify the pattern between engagement level and the topic discussed
- Based on the relationship between engagement level and picture tags
(iv) Identify the relationship between the performance of SGAG’s page and the topics discussed
- Based on the relationship between lifetime total likes (SGAG’s page) and frequently mentioned picture tags
(v) Identify the relationship between the composition of SGAG’s Facebook fans and the topics discussed
- Based on the relationship between the proportion of Male and Female fans from a range of age group and frequently mentioned picture tags
(vi) Identify time-series pattern (eg. the day and time posted) and cluster posts according to engagement level.
- Based on the date and time posted for every posts and the engagement level