Difference between revisions of "ANLY482 AY1516 G1 Team Skulptors - Methodology"

From Analytics Practicum
Jump to navigation Jump to search
(Created page with "<!--Main Navigation--> <center> 160px|link=ANLY482 AY1516 G1 Team Skulptors {|style="background-color:#ffffff; color:#000000; width="100%" cellspa...")
 
Line 31: Line 31:
 
<!--Content Start-->
 
<!--Content Start-->
 
<div align="left">
 
<div align="left">
−
<div style="background: #d9d9d9; padding: 12px; font-family: Impact; font-size: 18px; font-weight: bold; line-height: 1em; text-indent: 15px; border-left: #595959 solid 32px;"><font color="black">Methodology</font></div>
+
<div style="background: #d9d9d9; padding: 12px; font-family: Impact; font-size: 18px; font-weight: bold; line-height: 1em; text-indent: 15px; border-left: #595959 solid 32px;"><font color="black">Choice of visualization tools changes (Control Chart vs Tree maps)</font></div>
 
<br/>
 
<br/>
−
The following methodology will be implemented to perform analysis on the past year data provided by our sponsor for the two companies. Before we begin with our analysis, the data needs to be thoroughly examined and prepared to ensure there is no possibility of garbage-in-garbage-out.<br/>
+
<b>Before – Control chart of warehouse on landing page</b><br/>
 
<br/>
 
<br/>
−
 
+
The landing page of the dashboard was initially designed with a control chart instead of a tree map as sponsors were unfamiliar with tree maps and its usage. The control chart in a warehouse context is a graph used to study both warehouse inbound and outbound of SKUs over time. Data are plotted in a time order with a central line for the average, an upper line for the control limit (80% percentile), and a lower line for the lower control limit (20% percentile). These lines are determined by the data obtained from the sponsors.<br/>
−
<b>Data Preparation</b><br/>
+
[[Image:Skulptors-ControlChart.png|500px|link=]]<br/>
−
 
 
−
Before going into elementary explorations of the data, data cleaning is to be conducted to remove irrelevant data that should not be included in our analysis. Some types of reason for exclusion includes:
 
−
* Repetition – Double scanning happens many of a times in the company’s warehouse. As such, there can be repetition of rows in the past year data that may falsify our findings.
 
−
* Missing data – if there are any rows of incomplete data, the row will be excluded.
 
−
* Over-spilling – As our project focuses on inflow rate and outflow rate of the SKUs movement, we will exclude data there does not account for the start of the SKUs to the end of it.
 
−
 
 
−
After the initial preparation, we will further develop and sort our data by investigating for potential groupings that can facilitate ease into performing our visualizations and analysis. This will also assist us in the development of our regression model to figure out what type of classification, A, B or C should we give to a specific SKU. <br/>
 
−
<br/>
 
−
 
 
−
<b>Control Chart</b><br/>
 
−
[[Image:Skulptors-ControlSample.jpg|500px|link=]]<br/>
 
−
The control chart is a graph used to study how a process changes over time. Data are plotted in time order. A control chart always has a central line for the average, an upper line for the upper control limit and a lower line for the lower control limit. These lines are determined from historical data. By comparing current data to these lines, we can draw conclusions about whether the process variation is consistent (in control) or is unpredictable (out of control, affected by special causes of variation). <br/>
 
−
With respect to our project, these control charts will provide us with insights on whether certain movements of SKUs are out of control e.g. too little movements which results in wastage of space in the warehouse. It also simplifies the performance to be easily read by the human eye. Our sponsor emphasizes on sustainability of our solution, which will be further tackled by this in our dashboard.<br/>
 
−
<br/>
 
−
 
 
−
<b>Time series line chart</b><br/>
 
−
[[Image:Skulptors-TimeseriesSample.jpg|500px|link=]]<br/>
 
−
A time series is a sequence of data points, typically consisting of successive measurements made over a time interval. With relation to the inbound and outbound rate visualization for different SKUs, the time series line graph will cater to this purpose.<br/>
 
−
<br/>
 
−
 
 
−
<b>Treemap</b><br/>
 
−
[[Image:Skulptors-TreemapSample.jpg|500px|link=]]<br/>
 
−
A Treemap can convey our hierarchical data with 2 additional attributes via color and size, allowing us to dissect the relationship between the two. The size of each block represents the percentage of the warehouse utilized, and we can allow the user to fill in the colors with other indicators, such as the type of SKUs.<br/>
 
−
 
 
−
<br/>
 
−
 
 
−
<div align="left">
 
−
<div style="background: #d9d9d9; padding: 12px; font-family: Impact; font-size: 18px; font-weight: bold; line-height: 1em; text-indent: 15px; border-left: #595959 solid 32px;"><font color="black">Technology</font></div>
 
 
<br/>
 
<br/>
−
<b>D3.js</b><br/>
+
<b>After – Tree maps of warehouse on landing page</b><br/>
−
 
 
−
D3.js is a Javascript library. As our client request for as low cost as possible, D3.js will be a good option. That is because it can work on websites. Thus, our client will be able to see the data visualization without paying and installing any software. Another benefit of D3.js is its flexibility. It allows control over the final result. D3.js will be used for the visualizations that was mentioned above. <br/>
 
 
<br/>
 
<br/>
−
 
+
Upon advice from Prof, the team managed to convince the sponsors on the adoption of a tree map on the application’s landing page. A tree map is a method for displaying hierarchical data by using nested rectangles illustrated by different sizes and colors of squares. The control chart on the other hand, will be incorporated as a subset of the treemap. This meant that each click on a specific square on the treemap will brings the sponsor to a page depicting the outbound control chart for that specific put location.<br/>
−
<b>JMP</b><br/>
+
[[Image:Skulptors-Treemap.png|500px|link=]]<br/>
−
 
+
The rationales of placing the tree map at the landing page is as follows:
−
JMP is developed by the JMP business unit of SAS Institute. It is the tool of choice for data explorers in every industry. We will be using JMP to perform Control Chart Analysis for the inbound and outbound rate so that we can classify the SKUs into the ABC.<br/>
+
# Realistic and practical
 +
#* Putting a tree map at the landing page would be appropriate for a warehouse context as it can be used and designed in the form of the warehouse’s floorplan.  
 +
# High level view of warehouse utilization
 +
#* The different squares of a tree map can be used to depict the put away locations in the warehouse whereas the size of each squares can be used to show the total quantity of SKUs held within it. In addition, the colors of each square can also be used to signify over utilized or underutilization of a warehouse put away location. The control chart on the other hand, only shows micro view of the total outbound from the warehouse. 
 +
# Ease of problem identification
 +
#* With different color and size dimension correlated in some way with the tree structure, it allows users to spot patterns that would otherwise be difficult. For instance, numerous red squares can easily signify that there the warehouse is not optimally utilized.
 +
# Efficient use of space
 +
#* In a warehouse context where there are hundreds of SKUs and put away locations, tree map suits the criteria of legibly displaying hundreds of items on the screen simultaneously compared to other charts (Pie, bar charts).<br/>
 
<br/>
 
<br/>
  
−
<b>SAS Enterprise Miner</b><br/>
 
−
 
−
SAS Enterprise Miner is a software developed by SAS Institute. We may be using SAS Enterprise Miner to perform other needed analysis.<br/>
 
 
<br/>
 
<br/>
−
 
+
<font style="font-size:90%" color="#737373"><i>For the list of methodology & technology proposed during the proposal stage, please refer to: </i></font>[[ANLY482 AY1516 G1 Team Skulptors - Methodology & Technology |<font style="font-size:90%" color="#737373"><i><u>Team Skulptors - Methodology & Technology</u></i></font>]]
−
<b>Java & Bootstrap</b><br/>
 
−
 
 
−
Java will be used to develop the skeleton of the application while Bootstrap will be used to beautify the application. As our client wanted a sustainable analysis, we will need to do up a simple application whereby they can upload their data to be processed for analysis.<br/>
 
−
<br/>
 
−
 
 
−
<b>OpenShift</b><br/>
 
−
 
 
−
OpenShift is a product from Red Hat. It is an open-source platform as a service. It serves as a platform to launch our application for live deployment.<br/>
 
−
 
 
 
<br/>
 
<br/>
 
<!--Content End-->
 
<!--Content End-->

Revision as of 01:01, 28 February 2016

Skulptors-Logo.png

Skulptors-HomeIcon.png   HOME Skulptors-AboutIcon.png   ABOUT US Skulptors-OverviewIcon.png   PROJECT OVERVIEW Skulptors-DataIcon.png   DATA ANALYSIS Skulptors-ProjMgmtIcon.png   PROJECT MANAGEMENT Skulptors-DocIcon.png   DOCUMENTATION
Summary Description Methodology Technology Limitations & ROI


Choice of visualization tools changes (Control Chart vs Tree maps)


Before – Control chart of warehouse on landing page

The landing page of the dashboard was initially designed with a control chart instead of a tree map as sponsors were unfamiliar with tree maps and its usage. The control chart in a warehouse context is a graph used to study both warehouse inbound and outbound of SKUs over time. Data are plotted in a time order with a central line for the average, an upper line for the control limit (80% percentile), and a lower line for the lower control limit (20% percentile). These lines are determined by the data obtained from the sponsors.
Skulptors-ControlChart.png

After – Tree maps of warehouse on landing page

Upon advice from Prof, the team managed to convince the sponsors on the adoption of a tree map on the application’s landing page. A tree map is a method for displaying hierarchical data by using nested rectangles illustrated by different sizes and colors of squares. The control chart on the other hand, will be incorporated as a subset of the treemap. This meant that each click on a specific square on the treemap will brings the sponsor to a page depicting the outbound control chart for that specific put location.
Skulptors-Treemap.png
The rationales of placing the tree map at the landing page is as follows:

  1. Realistic and practical
    • Putting a tree map at the landing page would be appropriate for a warehouse context as it can be used and designed in the form of the warehouse’s floorplan.
  2. High level view of warehouse utilization
    • The different squares of a tree map can be used to depict the put away locations in the warehouse whereas the size of each squares can be used to show the total quantity of SKUs held within it. In addition, the colors of each square can also be used to signify over utilized or underutilization of a warehouse put away location. The control chart on the other hand, only shows micro view of the total outbound from the warehouse.
  3. Ease of problem identification
    • With different color and size dimension correlated in some way with the tree structure, it allows users to spot patterns that would otherwise be difficult. For instance, numerous red squares can easily signify that there the warehouse is not optimally utilized.
  4. Efficient use of space
    • In a warehouse context where there are hundreds of SKUs and put away locations, tree map suits the criteria of legibly displaying hundreds of items on the screen simultaneously compared to other charts (Pie, bar charts).



For the list of methodology & technology proposed during the proposal stage, please refer to: Team Skulptors - Methodology & Technology

Retrieved from ‘https://wiki.smu.edu.sg/ANLY482/index.php?title=ANLY482_AY1516_G1_Team_Skulptors_-_Methodology&oldid=5504’