Difference between revisions of "ANLY482 AY2017-18 T1 Group2 Project EZLin Scope"
(Created page with "<br> <font face="Arimo"> {| style="background-color:#FFFFFF; color:##228B22 padding: 5px 0 0 0;" width="100%" cellspacing="0" cellpadding="0" valign="top" border="0" | | styl...") |
|||
| Line 40: | Line 40: | ||
|} | |} | ||
<!--/Sub Header--> | <!--/Sub Header--> | ||
| + | |||
| + | <br/> | ||
| + | '''Phase 0: Understanding Data & Supply Chain Context''' | ||
| + | <br/> | ||
| + | Upon obtaining the raw data from the sponsor, we will be working, with the assistance of the client, to understand the data based on the client’s storage of information in its system. This includes mapping the entire supply chain flow and process of the client as well as seeking to understand what are the supply chain terminologies that the company uses. Upon doing so, we will proceed to add or edit variables/categories to add better context in allowing us to gain a better understanding of the data. | ||
| + | |||
| + | <br/> | ||
| + | '''Phase 1: Data Cleaning & Exploration''' | ||
| + | <br/> | ||
| + | Given the raw form of the data, we will need to clean the data before running any form of exploration or analysis. This includes but are not limited to the following steps, | ||
| + | * Checking for anomalies and outliers | ||
| + | * Deciding on action for missing data | ||
| + | * De-duplicating of any fields | ||
| + | * Standardising and normalising data | ||
| + | * Deciding and documenting on assumptions made. | ||
| + | Only after cleaning the dataset will we be able to do Exploratory Data Analysis (EDA) data set, which includes but are not limited to, | ||
| + | * Plotting the raw data | ||
| + | * Examining the distribution of the variables | ||
| + | * Studying the relationships between exploratory variables | ||
| + | * Conducting cross sectional and longitudinal analysis with the different factory locations | ||
| + | |||
| + | <br/> | ||
| + | '''Phase 2: Automating data retrieval process and modeling data for predictive analytics''' | ||
| + | <br/> | ||
| + | Once we have clearly understood the relationships and examined the variables in the dataset, we will proceed to develop a model for our client to firstly automate its data retrieval process. Given the current manual process, it’s important to be able to automate this retrieval and cleaning of data to allow the company to conduct its initial analysis. Afterwhich, we will seek to do modeling such that it will allow for predictive cost analytics. This will include external factors as mentioned above and developing as well as testing of hypotheses to determine the validity of our assumptions. | ||
| + | <br/> | ||
Revision as of 21:50, 25 August 2017
Phase 0: Understanding Data & Supply Chain Context
Upon obtaining the raw data from the sponsor, we will be working, with the assistance of the client, to understand the data based on the client’s storage of information in its system. This includes mapping the entire supply chain flow and process of the client as well as seeking to understand what are the supply chain terminologies that the company uses. Upon doing so, we will proceed to add or edit variables/categories to add better context in allowing us to gain a better understanding of the data.
Phase 1: Data Cleaning & Exploration
Given the raw form of the data, we will need to clean the data before running any form of exploration or analysis. This includes but are not limited to the following steps,
- Checking for anomalies and outliers
- Deciding on action for missing data
- De-duplicating of any fields
- Standardising and normalising data
- Deciding and documenting on assumptions made.
Only after cleaning the dataset will we be able to do Exploratory Data Analysis (EDA) data set, which includes but are not limited to,
- Plotting the raw data
- Examining the distribution of the variables
- Studying the relationships between exploratory variables
- Conducting cross sectional and longitudinal analysis with the different factory locations
Phase 2: Automating data retrieval process and modeling data for predictive analytics
Once we have clearly understood the relationships and examined the variables in the dataset, we will proceed to develop a model for our client to firstly automate its data retrieval process. Given the current manual process, it’s important to be able to automate this retrieval and cleaning of data to allow the company to conduct its initial analysis. Afterwhich, we will seek to do modeling such that it will allow for predictive cost analytics. This will include external factors as mentioned above and developing as well as testing of hypotheses to determine the validity of our assumptions.
