Data Warehousing

1
Data Warehousing
Virtual University of Pakistan

• Lecture-38
• Case Study Agri-Data Warehouse

Ahsan Abdullah Assoc. Prof. Head Center for
Agro-Informatics Research www.nu.edu.pk/cairindex.
asp FAST National University of Computers
Emerging Sciences, Islamabad
2
Step-5 Surprise case
Graphics
Sucking pests
Ball Worm Complex
SBW Spotted Ball Worm ABW Army Ball Worm PBW
Pink Ball Worm
If pest population is low, predator population
will also be low, because there will be less
food for predators to live on i.e. pests.
3
Step-6 Data Acquisition Cleansing
Hand filled pest scouting sheet
Graphics
Typed pest scouting sheet
4
Step-6 Issues
5
Step-6 Why the issues?
6
Step-7 Transform, Transport Populate
7
Motivation For Transformation
Graphics
8
Step-7 Resolving the issue
Graphics
9
Step-8 Middleware Connectivity
10
Step-9-11 Prototyping, Querying Reporting

• SELECT Date_of_Visit, AVG(Predators),
  AVG(Dose1Dose2Dose3Dose4)
• FROM Scouting_Data
• WHERE Date_of_Visit lt 12/31/2001 and predators
  gt 0
• GROUP BY Date_of_Visit

Graphics
11
Step-12 Deployment System Management
12
Agri-DSS usage Data Validation
13
Agri-DSS usage Data Validation Graph
ALL goes to graphics
14
Agri-DSS usage FAO report
15
Graph
Graphics
Why negative correlation between yield and
pesticides?
Using pesticides to increase yield.
16
Agri-DSS usage Spray Dates
17
Agri-DSS usage Spray Dates Graph
18
Agri-DSS usage Explaining Findings
19
Agri-DSS usage Sowing Dates
Graphics
20
Conclusions Lessons

• ETL is a big issue.
• Each farmer is repeatedly visited
• There is a skewness in the scouting data.
• Decision-making goes all the way down to the
  extension level.

All goes to graphics