From Visual Analytics for Business Intelligence
Jump to: navigation, search

Vertias Logo.png

Who are We?

We are:

  • Genevive Chan Keng Ling
  • Ma Myat Noe Mon
  • Ong Jing Yun

Our Project: The Secret Life of the FTSE 100


ActionAid has found that 98 out of 100 biggest groups listed on the London Stock Exchange actually use tax havens. The findings brought a major concern as multinational companies are using tax havens to the extent of relaxing the rules designed to prevent tax-haven abuse. The banks and financial sector are believed to be the heaviest users of tax havens despite the fact that they are largely profitable. This is disturbing since they are abusing the the deals of their tax bills. This has serious repercussions on the tax havens that are developing countries and who are struggling to improve their citizens standard of living.

ActionAid has generated a report that exposed more about the usage of tax havens by these British companies. However, the data and graph was revealed in a static manner. Veritas decided to look into the issue and figure out more about the usages of tax haven. We want to present the data in a more dynamic and intuitive manner to let audiences understand the gravity of tax haven usage by these companies.

We found a related piece of work about Tax Haven usage by FTSE 100 companies click. It shows the proportion of companies and sectors that use tax havens. We find it a fairly useful example to refer to, but we added in extra information to our visualisation that we hope will add value to exposing the issue to audiences.

Click here to view ActionAid's report on Tax Havens.

Follow this link to learn more about ActionAid's campaign to help developing countries against tax dodgers & know how you can help.

Problem & Motivation

This project aims to use data visualisation to discover and expose which FTSE 100 companies/sectors use tax havens to avoid paying taxes and to show the inequity caused to the vulnerable involved.

It is true that many of us do not like the idea of paying taxes. The problem now is that many are avoiding the taxes and it's a legal practice for many multinationals. However, when these countries exploit the poor to pick up the bill so as to minimise their taxes, it can desperately destroy the growth of developing countries.

Therefore, taxes can affect both rich and poor. Our motivation will be finding out who is over-using the tax haven and who might be most vulnerable.


Visualisation 1 Visualisation 2
The first visualisation is a Treemap. The treemap shows the proportion of sectors in tax havens that are either developing or developed countries. It allows users to see which sector are prevalent as tax haven users countries with High-income, Upper-middle-income, Lower-middle-income & Low-income. The different colours represent the different industries and companies that use tax havens.

In this visualisation, we can see compare the usage by each sector/company of the tax havens in countries with different income level. In other words, we can see who is the main culprit of tax haven abuse in income level. The number of subsidiary companies by each parent company using tax havens is printed within the treemap cells too.

The second visualisation is a bar chart. The bar chart shows the exact number of companies using tax havens by each sector and drilled down to show how many operations of each company use tax havens to which countries they are using as tax havens.

In this visualisation, we can see more specific details of the usage of each sector and companies.

Technology used

  • D3.js

More about the storyboard and technologies can be found in our research paper.


Data Collection and Preparation Process

The data we used is available at this link. Here are the steps we took to process the data for visualisation.

  1. We deleted the records with subsidiary companies that do not use tax havens, i.e. the records with tax haven value 0. We also deleted records with income classification as "Unclassified".
  2. Then we added in information about whether the specific tax haven country is a developed country or a developing country by checking the Trademap Organisation website. Additional research had to be done for certain countries because they were not included in the website. We used JMP Pro join funtion to add in the economic status of each tax haven country.
  3. Finally, we had to prepare the data in the appropriate JSON format for the various treemap visualisations we tried out. The treemap required a nested JSON hierarchy. We tried looking for online tools to help transform the flat data format our original data was in to the nest JSON one, but we looked to no avail. We then tried to write a Java program to put the data in the appropriate structure, but found the debugging of the code too time-consuming. We finally decided to use JMP Pro and Microsoft Excel to transform the data to the correct format.

More about the ETL process can be found in our research paper.

Project Scope

We want to build an application provides visually friendly infographics to expose the usage of Tax Havens by companies.

The following are visualisations we want to show:

  1. Analyse Involvements of Countries Concerned & their Economic status
  2. Analyse Sectors & Companies Extent of Tax Haven Usage

Work Allocation

Week Milestone (By End of Specific Week) In-Charge
  • Complete ETL Process
  • Confirm which Technology to use
  • Update Wiki page with Proposal
  • Come up with Prototype
  • All
  • Learn how to use D3.js
  • All
  • Dashboard 1: Bar Graph
  • Dashboard 2: Tree Map
  • Noe
  • Ginny & Genevive
  • Check & Test both Dashboard 1 & 2
  • Ginny
  • Prepare Poster
  • Update Wiki page
  • Ginny
  • Ginny & Genevive
  • Poster Presentation
  • Complete Final Report
  • Submit deliverables

Project Risk

Key Technical Challenges & Approach To Address Challenge

Challenge Approach
ETL process was tedious because of the need to arrange the data into the appropriate JSON format. Different methods of visualisations required different JSON formats. We experimented with different treemap visualisation and, thus, needed to rearrange the data several times. We used JMP Pro and excel to facilitate the ETL process. We also used Text Fixer [1] to remove tabs and white spaces. Another online tool we found useful was JSONLint [2], which is a tool to check if our JSON file is in the right format.
Learning how to use D3.js to portray useful information(Limitation: Time Constraints)
  • Peer coaching in team and across teams was helpful.
  • Tutorial self-learn
  • d3.js google group discussions were especially helpful. Peter Rust and Mike Bostock of d3.js group were helpful too.
Specific to Application: Import function & make it Flexible to Accept New Data Research how to use CSV import function. However, we have decided to drop the function due to time constraints and technical challenges.

Click here to view the Project Poster.


Related Works

Here are some related works in terms of data visualization type we can refer to when developing this application. They are :-

1. Flex Dependency Graph for course network

2. Sample Network Visualization

3. Drill Down Graphs

4. Topic related visualisation

Previous Idea

Initially our team decided to focus on Boss Bidding. We wanted to find out the trend of the Boss Bidding. We wanted to provide some assistance for students who wish to know more about the bidding trends clearly and help them estimate the bidding costs for their respective modules. Unfortunately due to some data limitations, we are unable to continue. Click here to view our previous proposal.


Please feel free to give helpful comments. Thank you.
Cite error: <ref> tags exist, but no <references/> tag was found