Project Work Component (CBAS) Mar 2014 PT Batch

50 %
50 %
Information about Project Work Component (CBAS) Mar 2014 PT Batch

Published on June 21, 2016

Author: ananaub

Source: slideshare.net

1. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   1 Name : Alvin Sim Buan Heng Submission Date : 5 April 2014 PART 1: DATA/INFORMATION ARCHITECTURE 1. AS IS Pain Points to be addressed: • Data-reporting speed • Data integrity • No insights of trend & analytics from their data 2. Explanation of Model Components: Component Explanation Data Warehouse Centrally managed and integrated database containing data from organization’s operational sources. In the diagram above, the Backbone System, Core System and Opportunity System provide the operational sources. Data Marts Subsections of a Data Warehouse. Extract, Transform & Load (ETL) Process Extract is the process of reading data from databases such as the corporate HR and Finance databases Transform is the process of converting extracted data from previous form into a form it needs to be in so as to be placed into another database. This process occurs using rules or lookup tables or combining data with other data. Load is the process of writing data into target database.

2. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   2 Component Explanation Business Intelligence Information (such as ROI Reporting) used by an enterprise to make decision. 3. Assumptions of the proposed model: • Data is 100% cleansed or validated during the ETL process • Error-free architecture PART 2: CRISP/DM PHASE 1 BUSINESS UNDERSTANDING a. Business Background • Best Car Dealer has enough data to run all Analytic Reports and Performance Indicators using acquired Analytics tools. • Company CIO was keen to identify trends from the data but the company has no expertise to do so. b. Determine Data Mining Goals • Observe for Warranty Claims pattern. PHASE 2 DATA UNDERSTANDING a. Collect Initial Data Report • Import data to RapidMiner (for Mac OS) b. Describe Data • Document all the columns in the Training Data and description of each Phase 1: Business Understanding   Phase 2: Data Understanding   Phase 3: Data Preparation  

3. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   3 c. Explore Data: Studying the Meta Data from Rapid Miner

4. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   4 d. Data Quality: Note down all the data inconsistency, missing data etc. in this report S/N Column Missing Data Data Inconsistency Description Recommendation 1 Workshop 1 Nominal value C has highest absolute count of 42, followed by B (36) and A (31). 1 count each for AB, BB and E. Fill missing data with C. Replace inconsistent data (AB, BB and E) with C 2 Model 1 Toyota (44), Mazda (36), BMW (30), BMWS(1), Proton (1). Fill missing data with Toyota. Change BMWS to BMW. 3 Causes 2 Cable (40), Mileage (37), Gear (32), Gears (1), Mileages (1) Fill missing data with Cable. Change Gears to Gear. Change Mileages to Mileage. 4 Comments 3 Inputs appear arbitrary and it is unlikely for one to derive any insights from this data. Remove this column since it is unlikely for this data set to provide insights into warranty claim trending analysis. 5 Claimed 1 Y(55), N(52), O (1), 1 (1), B (1), False (1), True (1) Replace all inconsistent data with Y except ‘False’ (replace with ‘N’) Fill missing data with Y 6 Accident 0 Y(59), N(50), Yes (2), No (2) Replace ‘Yes’ with ‘Y’ and ‘No’ with ‘N’ 7 Customer, Custid & Date 0 NA Remove these columns since it is unlikely for these data sets to provide insights into warranty claim trending analysis.

5. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   5 PHASE 3 DATA PREPARATION All the phases as needed whatever operators you use in Rapid Miner for Data Cleansing/scrubbing document in each of the phase

6. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   6 PART 3: MODELING IN RAPIDMINER AND THE PHASES Creating the model in RapidMiner and export as the .rmp file PHASE 4 MODELING WBS - Select Modeling Technique: Which model you choose, why and assumptions if any. S/N Column Missing Data Data Inconsistency Description Recommendations Assumptions Selected Modeling Technique 1 Workshop 1 Nominal value C has highest absolute count of 42, followed by B (36) and A (31). 1 count each for AB, BB and E. Fill missing data with C. Replace inconsistent data (AB, BB and E) with C Replacing missing and inconsistent data with the majority is not going to profoundly impact the outcome. Replace Replace Missing Values 2 Model 1 Toyota (44), Mazda (36), BMW (30), BMWS(1), Proton (1). Fill missing data with Toyota. Change BMWS to BMW. Replacing missing data with the majority is not going to profoundly impact the outcome. Spelling error (BMWS). Replace Replace Missing Values 3 Causes 2 Cable (40), Mileage (37), Gear (32), Gears (1), Mileages (1) Fill missing data with Cable. Change Gears to Gear. Change Mileages to Mileage. Replacing missing data with the majority is not going to profoundly impact the outcome. Grammar error (Gears, Mileages). Replace Replace Missing Values 4 Comments 3 Inputs appear arbitrary and it is unlikely for one to derive any insights from this data. Remove this column Unlikely for this data set to add value into warranty claim trending analysis. Select Attributes 5 Claimed 1 Y(55), N(52), O (1), 1 (1), B (1), False (1), True (1) Replace all inconsistent data with Y except ‘False’ (replace with ‘N’) Fill missing data with Y Replacing missing and inconsistent data with the majority is not going to profoundly impact the outcome. ‘True’ refers to ‘Y’ and ‘False’ refers to ‘N’. Replace Replace Missing Values 6 Accident 0 Y(59), N(50), Yes (2), No (2) Replace ‘Yes’ with ‘Y’ and ‘No’ with ‘N’ ‘Yes’ refers to ‘Y’ and ‘No’ refers to ‘N’. Replace 7 Customer, Custid 0 NA Remove both columns. Unlikely for this data set to add value into warranty claim trending analysis. Select Attributes

7. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   7 PHASE 5 EVALUATION Evaluate the results: Write a use case for the business what is the outcome if any of the Data Mining From Data Mining, the client may be keen to know the following insights: 1. Customers’ Choice: Which is the most preferred workshop? 2. Logistics Allocation: What are the car model spare parts each workshop should have in stock? 3. Manpower Deployment: How to deploy technicians? S/N OPERATOR OBSERVATIONS INTERPRETATIONS 1 Correlation Strongest Correlation Coefficient: • Claimed vs Accident (0.911) Weakest Correlation Coefficient: • Model vs Claimed (-0.117) It is a given that customers will make accident claims. However, the client may wish to note that the number of accident claims is independent from the car models driven by the claimants. 2 Association 3 K- Means Clustering Referencing from Cluster 0, 3 & 4 only: BMW drivers seem to prefer Workshop B to solve their mileage issues. Mazda drivers seem to prefer Workshop C to solve their cable issues. Toyota drivers seem to prefer Workshop A & C to solve their gear issues. Workshop A & C seem to be the preferred workshops for Japanese car drivers.

8. Project Work Component (CBAS) Mar 2014 PT Batch Alvin Sim   8 PHASE 6 DEPLOYMENT Plan Deployment: How you plan to deploy this in your organization. In the coming Financial Year, Best Car Dealer may wish to stock up their Workshops A, B & C with certain spare parts and to deploy competent technicians specialized in accident causes based on the following recommendations: Workshop Model Parts Causes A Toyota Gear B BMW Mileage C Toyota & Mazda Gear & Cable

Add a comment

Related pages

Google

Advertising Programmes Business Solutions +Google About Google Google.com © 2016 - Privacy - Terms. Search; Images; Maps; Play; YouTube; News; Gmail ...
Read more

2014 02 20 quedancor daily news monitor by PIO ... - issuu

Issuu is a digital publishing platform that makes it simple to publish magazines, ... 2014 02 20 quedancor daily news monitor, Author: PIO QUEDANCOR, ...
Read more

ERP - scribd.com

Customer relationship management Supply chain management Project management Access control ... based costing. work ... component/ content/ article ...
Read more

The Transformation of the Workplace Through ... - JD Supra

The Transformation of the Workplace Through Robotics, Artificial Intelligence, and Automation - February 2014 . ... 2014 Roundtable on The Future of the ...
Read more

Pages 10 - Www.terkud.com

book laptop repair complete guide including motherboard component level ... symposium ipec 2014 wroclaw poland ... sequencing batch reactor ...
Read more

Full text of "Report of program activities : National ...

American Libraries Canadian Libraries Universal Library Shareware CD-ROMs Community Texts Project Gutenberg ... of program activities : National Cancer ...
Read more

52.225-17

52.242-15 Stop-Work Order AUG 1989 ... For items that are serialized within the original part, lot, or batch ... (MAR 2014) (a) Performance ...
Read more

META-INF/MANIFEST.MFname/audet/samuel/shorttyping

btch,batch bhes,branches blde,blonde ... coms,comments cntc,contact cion,collection ... cmnt,component curs,curious cmes,communities
Read more