GOOD PRACTICES FOR DATA DISSEMINATION Juraj Riecan Director, UN-ESCWA Statistics Division
Overview of dissemination practices Traditional databases Front-ending Process-oriented architecture From recordsets to data cubes
Current dissemination practices Static file formats Spreadsheets National databases
√ X > Static files Fully controlled presentation Combine data, text, charts Quick, easy, cheap √ Only a snapshot Data not downloadable Manual updates X Key indicators Highlights >
√ X > Spreadsheets Easy format for data analysis Downloadable, Customizable √ Users have to download all Manual updates X Data files for further analysis >
√ X > National databases Customized queries Flexible downloads Easy updates √ Important implementation costs IT infrastructure System maintenance X General data dissemination >
Traditional databases
Front-ending Database 3 Database 2 Database 1
Process oriented architecture Production Database Database 3 Dissemination Database Database 2 Database 1
Statbank SuperWEB DevInfo New Chronos Beyond 20/20 PC-Axis OECD.Stat Dissemination Database DevInfo New Chronos Beyond 20/20 PC-Axis OECD.Stat
From recordsets to hypercubes Relational databases Hypercubes
Evolution from paper records Record = data item & metadata Recordsets Evolution from paper records Sequential reading Record = data item & metadata
Standard Query Language (SQL) Generic database tools Relational Databases Reporting country Trade Volume Importing country Flexibility Exporting country Standard Query Language (SQL) Generic database tools
Data Cube (Hypercube) Age Country 2001 2002 2003 2004 Year 65+ 16-64 0-15 Country C Country Country B Country A 2001 2002 2003 2004 Year
Combining multiple classification Easy cross-tabulation Data Cube (Hypercube) √ Most common approach √ Flexible √ Expandable √ Combining multiple classification √ Easy cross-tabulation
Conclusion – Summary Process oriented Architecture Dissemination databases Hypercube data model
THANK YOU riecan@un.org