Académique Documents
Professionnel Documents
Culture Documents
Background
Database
We have a Postgres database of about 60 million betting transactions, which we would wish
to analyze by means of an OLAP cube.
Cubes
1. A Transactions cube, where the fact table is the transactions table (60 million rows).
2. A Payments cube, where the fact table is the payments table (1 million rows).
3. A Clients cube, where the fact table is the client table (250,000 rows).
Transactions Cube
Measures:
– turnover
– profit/loss
Profit/loss is the difference between the price a bet is purchased and a price the bet is then
sold/expired.
Dimensions:
LEVEL1 LEVEL2 LEVEL3 LEVEL4
One Touch
No Touch
Path Dependent …
Bet type
Not path dependent Rise/Fall variable return
Rise/Fall fixed return
Higher Lower
Ends in/out
…
FTSE
Stock indices …
USD/JPY
FX GBP/USD
…
Commodities Gold
....
Random 50
Random indices Random indices …
1 day
< 1 week 2 days
3 days
Daily 4+ days
1-2 weeks
2-4 weeks
> 1 week 4-8 weeks
8+ weeks
Country Canada
Affiliate referrer Mexico
…
Americas
Virtual France
Europe ..
Asia Singapore
Client Country ..
Africa
Nigeria
< 25 years old …
Age 25-40 years old
Real 40+ years old
Authentication status
A/C status (disabled etc)
Affiliate referrer < 1 week
Days since a/c opened 1 week-1 month
.... 1-3 months
3-6 months
6+ months
Weeks
Time Years Quarters Day of week
Months
Measures:
– deposits
– withdrawals
Dimensions:
Weeks
Time Years Quarters Day of week
Months
Country Canada
Affiliate referrer Mexico
…
Americas
Virtual France
Europe ..
Asia Singapore
Client Country ..
Africa
Nigeria
< 25 years old …
Age 25-40 years old
Real 40+ years old
Authentication status
A/C status (disabled etc)
Affiliate referrer < 1 week
Days since a/c opened 1 week-1 month
.... 1-3 months
3-6 months
6+ months
Barclays
Lloyds
Moneybookers
Credit card ....
Envoy
Payment Bank wire Direct
....
Moneybookers
E-wallet Neteller
....
Western Union
Other ....
Client Cube
Measures:
TBD
Dimensions:
TBD