Group06 proposal

Growth Signal

Proposal

Poster

Application

Research Paper

Problem & Motivation

Problem

In Singapore, education is one of the sectors that is heavily invested by the state besides defence and health. As years progress, more data are accumulated and made available to the public by various government institutions such as Singapore Department of Statistics and data.gov. Despite the abundance of data, it is often not easy for someone fresh out of post-secondary education to plan for further studies in an efficient and thorough manner due to decentralisation and lack of visualisation (data are in raw/unstructured/tabular form).

Motivation

Enrolling in higher education such as polytechnics, especially universities is becoming a commonplace in Singapore. Not only does it prepare the next generation for the future (individual level), the country as a whole can significantly benefit as a result of a highly-educated population. This interactive visualisation will serve as a platform for post-secondary individuals to analyse the changing landscape and trends of higher education where they can uncover interesting insights and make well-informed decisions to ensure a sustainable future.

Objectives

Target Group: Post-Secondary individuals

With abundance of data and the lack of insights, our team would create dashboards to enable post-secondary individuals to answer the following questions:

• What are the enrolment rates of the different institutions? Are they increasing or decreasing?
• How much are the Singapore government spending on education throughout the years?
• Which degree pays most?
• What are the employment rates for the various courses?

Dataset

Data/Source	Variables/Description	Rationale & Methodology
Google Analytics Data from Client's Account (1 Jul 2016 to 1 Mar 2020) Livestream capable	Users New Users Sessions Number of Sessions Per User Page Views Pages / Session Average Session Per User Page Views Pages / Session Avg Session Duration Bounce Rate Conversion values Number of Users all at Stages of Customer Journey	This is a clickstream dataset which provides strategic data from all traffic sources. Non exhaustive list.

Background Survey

Research	Learning points	Improvement
Source: https://www.tapclicks.com/resources/blog/6-steps-marketing-dashboard/	Example of sales funnel (top right) on the to show conversion of Users to Leads and Trial. Can be broken down into timeline as well. The sales funnel would give an accurate idea of how many percentage of users are really interested in the advertisement or those that actually manage to secure leads. However, diagrams in 3D can be misleading due to overlaps and is not very neat. How many leads in the last 90 days or this month can be combined.	Can add Interaction for across a time period. Funnel can be broken down into more categories such as number of potential leads, marketing leads, sales leads, deals etc. Can give top 5 sources of traffic
Source:https://www.klipfolio.com/resources/dashboard-examples/marketing/digital-marketing-dashboard	Another dashboard example provided by kilpfolio. Useful example to show metrics such as quality of leads and conversion metrics, leads by region and top few social media channels. It can be animated or add a scale to show changes across timing However, leads by region is too small. Google web analytics Web traffic targets is not labeled and just shows the bar. Can change to label with highlight of red and green (based on whether the progress is negative or positive)	Perhaps can show the top 5 areas instead and table can be replaced by a bar graph. Key conversion metrics and quality of leads can be replaced with line graph to show changes over time.

References

• 6 Steps to Creating a Better Marketing Dashboard
• Klipfolio - Digital Marketing Dashboard
• 100 Tasks
• Web Analytics 2.0 by Avinash Kaushik

Technical Challenges

Challenge	Description	Mitigation Plan
Unfamiliarity with Tools	Most of the team members are not too familiar with R and R Shiny.	Make use of external sources for information and troubleshooting Utilise DataCamp to learn, practice and brush up skills on R and R Shiny
Lack of Marketing Domain Expertise	There may be gaps in terms of marketing jargons and concepts we are unaware of or fail to pick up in such a short time span.	Learn and apply the best practices from Web Analytics 2.0 by Avinash Kaushik Consult with client for feedback on the dashboard
Data Quality & Integrity	As Google Analytics uses clickstream data, there exist a possibility of bot traffic registering as data points. Much time is also needed to clean the data before feeding it into the visualisation.	Conduct proper EDA and data cleaning methods to minimise quality problems Apply filters and follow best practices from the industry to reduce data abnormalities

Storyboard

Dashboards	Description
Dashboard 1: Key Metrics Performance	Key metrics performance indicators would provide executives on the pulse of the business. A dropdown box at the top left would allow the selection of time period and metric the graph would represent. Instead of providing a value for the metric per time period, Sparklines and percentage changes from the previous time period would show how the metric changed.
Dashboard 2: Sales Funnel Analysis with Segementation	Sales funnel analysis would allow executives to show volume of users at each part of the customer journey. Percentage changes would show the difference between time period, this would allow executives to identify which part of the sales funnel is generating the most friction for buyers. Upon clicking or selecting from a dropdown menu, there would be options to filter via traffic source (Organic, Direct, Paid Social etc). This would allow executives to determine the best traffic source that leads to sales and optimise their marketing spend. Our team is still exploring the possibility of adding more filters such as geographic location or demographics to allow more insights.
Dashboard 3: Customer Acquisition Cost vs Customer Lifetime Value	The last dashboard would be the show the cornerstone of all businesses. Would their business survive in the long run? Are they spending too much to acquire a customer? How long would it take for their customer acquisition cost to break even? Some calculation for context: Customer Acquisition Cost (CAC): How much does it cost to acquire a customer CAC = Total Acquisition Cost / # of Customers Acquired Customer Lifetime Value (CLV): How much does a customer bring over the lifetime CLV = Total Acquisition Cost / # of Customers Acquired Filters per time period and channel would allow further segmentation.

Project Timeline & Milestones

Detailed Project Timeline from Asana (team's task tracker)

Comments

Name	Date	Comments
(Name)	(Date)	(Comment)
(Name)	(Date)	(Comment)
(Name)	(Date)	(Comment)