Transcript Schonfeldt_Online M&E for Nationwide HIV Initiative_loveLife

Scope
Approximately 32 monitoring tools and 27
datasets
Largest dataset - approximately 1,2 million
entries
Problematic quality assurance 32 data
capturing regions, especially deep rural
Challenges of the Old System
Thousands of social, demographic,
trend, GIS and relational data in
various forms and schemas
Constant data cleaning, high level of
duplication, inconsistent references
and orphaned data
Data was emailed, faxed, sent by
pigeon ;)
Enter Third Party Developers
Forcing predefined business models onto
social data systems – requires a mind shift
Third party issues regarding data and
loveLife structure – turn around time,
quality of backend
Technology and Costs –fixed diginet costs,
deep rural areas
2006 – Taking control
New provincial structures with 9 Provincial
Administrators capturing data online
Instantaneous access to information and automated
reports/graphs via 3G routers
5 Point data Quality Check from site to provincial level
Final quality check at national level
Site level training
Comprehensive monitoring manuals and multi level
support
Accessible via loveLife VPN, utilising LAN
infrastructure; in addition to harnessing the benefits of
emerging GPRS, 3G and HSDPA connectivity = huge
cost benefit
Development Process
Project Definition Workshop
Defined objectives, dependencies, assumptions and
constraints; functional spec, project scope, prioritised
feature lists, design and quality goals, system
architecture, logical and physical design
Project lifecycle - waterfall approach for each phase –
each phase –cyclic approach
Forums throughout development:
Internal Status Meetings (ISM)
Project Status Meetings (PSM)
Action and Issue Logs
Technology
Visual Studio 2005, JavaScript, AJAX
(simplifying navigation)
ASP.NET (richer interface)
MS SQL Server 2000
Crystal Reports XI Developer Edition
dotNet Charting
CAN BE DONE WITH OPEN SOURCE
Best Practice Functionality:
Avoid Performance Bottlenecks:
Client side validation
Make use of pure html where possible
Page refresh rates
User Ability:
Using master pages ensures a consistent look and
feel
Required versus operational data
Performance:
Use stored procedures effectively (code)
Security:
Roles and Access E.g. administrator versus
provincial users
Data encryption
The current online database
Using CrystalReportViewer object model,
logging directly into secure SQL Server Db
= interactive reports
Display of high performance charting control
technologies, aggregating systems with
calculation support (.netCHARTING web
forms)
Displays massive amounts of dynamically
generated data on front-end
Dynamically generated grid views populated
by data sources - created by parameter
driven stored procedures
Export to Excel, PDF and RTF
Current Challenges
Continuous changes around municipal
boundaries
Census 2001 data not aligned to
current boundaries
Infrastructure in South Africa
Changing Expectations
Underreporting – incomplete data sets
Contributions
Up to date and extensive GIS
coordinate database (public clinics and
schools)
Infrastructure
Volunteer Recognition (Incentives)
Accurately linked data
Front end Reporting
Data as management tool
Cost effective
Future Plans
More sophisticated data analysis
Increased number of publications and
maps
Continual improvement of tools,
datasets and database
Improved turn around time
100% Quality
Continued Maintenance of back ends