ANLY482 AY2016-17 T1 Group5 - Project Description

Overview

Description

Methodology

Visualizations

Technology

Limitations

Motivation

1. Traditional POS systems are used mainly for completing orders and payments from a terminal. Data collected from these systems are hence limited to sales figures. A successful F&B business should not only analyse sales figures but also other aspects of the business such as efficiency of the current operations, inventory stocking, marketing campaigns and etc.
2. As HoiPOS systems are programmed to offer a broad array of features such as sending/receiving orders to/from the kitchen, modifying orders, completing orders, and completing different modes of payments for an order. The data that is collected from these functions are more comprehensive compared to data from traditional POS systems. By analyzing and making sense out of these data, it could help businesses identify trend and find out information that is unknown to them and thus helping them improve, profit and efficiency. While the current analysis is helpful as it compares historic data and real-time data to compare trends and identify pitfalls and successes in the business, the visualizations used can be improved to offer a more comprehensive understanding of the data.
3. Most of HoiPOS competitors do not offer visualizations of data that is actually useful to businesses. Merely identifying that a business has performed better last week than this week does not provide useful information to the owners. Knowing the cause of last week’s better performance will be more useful as they can replicate the cause to achieve success again. An example would be identifying which particular item’s popularity is contributing to the better sales performance in the business allows the owner to inform the kitchen to prepare a certain dish at a faster rate will translate to higher table turnover rate.
4. Furthermore, most of HoiPOS competitors do not offer comprehensive visualizations of the data analysis which often confuses the client and hence not maximizing the potential of the information extracted from the data.

Objective

The objective of this project is to provide an interactive dashboard of different visualizations of data collected from HoiPOS systems. The project aims to improve HoiPOS’ data visualizations as well as add more visualizations of data analysis to value-add their POS system to clients. This project also aims to be able to apply descriptive analytics and effective visualizations to gain insights not only on the sales performance but also operations and marketing campaigns.

The final product will consists of an interactive visualization dashboard which will be designed to provide visualizations that can be easily understood by the layman as well. Intuitive functions such as cross filtering, will be implemented to allow business owners to conduct exploratory data analysis without needing technical knowledge. To ensure that the dashboard is easy to use, the team will also conduct several user testing with stakeholders.

The main objective of the project would be to develop the following functions/visualizations:

• Filter data by date/time
  • This allows to choose certain periods of the year to analyse as different seasons of the year could have a higher anticipated sales performance e.g restaurants near F1 location during F1 event
  • Being able to view analysis of data at a more micro level also allows the business owner to understand which time periods are peak periods and hence make better human resource deployment decisions quickly
  • Anomalies in the data can also be noticed quickly and will be able to have a reference point of why the anomaly happened using the date and time. E.g Spike in orders during F1 event

• Dashboard for visualizations
  • Sales Performance over time:
    • Includes overview of sales, average amount earned a day, number of orders, target analysis and more
  • Compare historic data and real-time data to notice gaps and/or trends

• Association Analysis of items
  • To allow users to identify popular combination of items. This information is useful when crafting marketing campaigns e.g combo meals of the popular combinations of items

• Visualization of Popular Items
  • To allow users to identify most popular items and/or identify items that are least popular. This information informs users about items that are not being ordered frequently could be removed from the menu to reduce wastage and costs
  • Exploring further for Popular Items - Visualization of Popular Items Modifier
    • To allow users to identify the most popular item modifier i.e hot/cold options for a cup of tea. This information may prompt users to modify the menu to accommodate for the popular preferences.
      • Users will be able to cross-filter the data using time as a variable to see the popular items at different time periods

• Visualization of Productivity Analysis
  • Users will be able to analyse time taken to prepare different food items and how it varies across the day. This insight will be able to aid business owners in making operational decisions with regards to human resource and productivity in the kitchen.

• Cross-filtering
  • Allow users to quickly have a glance of different set variables across a few different visualizations.

Provided Data

The data are provided by the sponsor in CSV format. The data consist of information of the order such as the staff in charge, order time, purchased item, promotion, discount, mode of payment. The required data will be extracted out and further processing will be done.

Note: Due to NDA, the company whose data were collected will not be disclosed and we will be using PLU code to represent the product.