ANLY482 AY2016-17 T2 Group22: Project Overview
|
|
|
|
|
|
Project Aim
In this project our team will be working closely with the customer insights department of Binge to realise the potential of their existing data to better differentiate and understand the various segment of their customers who use their live betting platform. This will allow the business to enhance the understanding of its customers, improve gameplay with the introduction of new bet types, and to improve its revenue and resource allocation.
Business Problem
Being relatively new to the Internet space, especially for live betting, Binge hasn’t been able to do any conclusive analysis of this data. They have many users who perform live transactions, pre-match transactions or a combination of both. To fully reap the benefit of being online and collecting all this valuable data about their customers, Binge needs a system that helps them analyse this data and get consumer insights about the trends and patterns in live transactions.
Motivation
1.Understanding Customers Behaviour The ultimate goal of data analysis in this case is to get actionable insights that can help us understand customer behaviour in different stages and different events for live transactions. This will help Binge get a holistic picture of the carrying types of customer and their behaviours. With that in mind, they can alter their target market depending on the match or stage of the match. 2.Customer Profiling Put the increase rise to understand the customers Understanding its customer is an important step Customer insights
Our project is largely focussing on helping our client understand the different customer behaviours in a live transaction setting. We aim to gather insights after analysing the data and creating different customer profiles for different types of behaviours. This will help in understanding its customers better and can work with some level of certainty even in a live setting.
Scope
The scope of our project is listed as follows: ¥ Data collection – this will be assisted by Binge due to the sensitivity of the data ¥ Data preparation – to ensure the data is clean and can be analysed using analytical software, and to recode variables for the program to read ¥ Model planning – to attach suitable models from literature reviews to the data set and to improve upon the model for this set of data ¥ Model analysis – to carry out the analysis of customer behaviour and to profile the customers into segments ¥ Model validation and implementation – to ensure the results are similar when attached to a different year of study and to document implementation procedures ¥ Model refinement – this will be assisted by Binge to obtain feedback and experience to improve the existing model ¥ Recommendations The project will mainly revolve around achieving the below mentioned objective. Our sponsor has given us the flexibility to read into the available data as deeply as we would like to, and add to the predetermined scope.
Data
We have managed to obtain 2 years of sports betting data – fiscal year 14/15 and 15/16 – collected from Binge online betting platform. The data that we have received is well documented and relatively clean, as such data clean-up would be kept at a minimal. Below is a table of the list of variables and the associated description.
Methodology
1.Model planning
Cluster Analysis To find out more about the segment of customers we hope to employ cluster analysis as a tool to partition customers. Data to use would include their bet frequency, transaction amount the amount of risk they take from the average value of their odds.
Decision Tree Analysis To find out the next step of action for a customer, a decision tree analysis can help predict the next step of the customer. This will allow us to analyse the probability of the customer proceeding to place the next bet and a sense of the amount the customer would be likely to place.
Model validation Finally, we will be verifying this analysis with a different fiscal year to ensure that the results do not differ significantly. If possible, we would be using a t-test to ensure that the differences are statistically insignificant.
2. Software The sponsor has granted us the flexibility to use any software tools suitable for the project. Hence, as a start we will be using tools that we are experience in such as JMP Pro and SAS Enterprise Miner for the analysis and modelling of the project.
