About the Solution/Project
This was a learning project I undertook to test my understanding of design concepts in BOARD. By using the Canadian census data, I wanted to see test the geographic tools in BOARD. This dataset has almost 490K datapoints with lat/long coordinates and complex aggregations to define many types of geographic regions. At the lowest level, we have one city block with the population, private dwellings and private dwellings occupied.
Video Walkthrough
To see a video walkthrough of the solution, visit this link on YouTube. --> Exploring Census Data Using the BOARD GeoMap Tool - YouTube
Takeaways
After experimenting with many approaches, the final solution is pretty simple. Here are some takeaways I learned.
- Use relationships to define all the geographic relations between entities
- Load all data at the same, lowest grain
- Add versions to improve performance of each screen
Attachments/Print Screens/Videos
- Navigation screen to zoom-in/zoom-out with the mouse
- Filter to specific geographic regions with pagers and select tools
- Entities
- Relationships - Starting with one city block, geographic groupings are very complex
- Cubes - Population is the main cube with many versions to ensure good performance.
- Data Readers - Just two data readers. One for metadata. One for facts.
Links to Datasources
Canadian Geographic Attributes File : http://www5.statcan.gc.ca/olc-cel/olc.action?objId=92-151-X&ObjType=2&lang=en&limit=0
Canadian Census Data Measures: Download, Census Profile, 2016 Census
Attachments
The Census.zip attachment includes the following items
- 2016_92-151_XBB.csv - This is the source CSV file used for both data readers
- Census.cpsx - The capsule file
- Census.hbmp - The database
If you want to play around with this solution, the archive should have all the components you need to do so. If you have any tips or suggestions for other design approaches, please reply so we can all learn.
Bob, Thank you for a great video and for demonstrating your takeaways. I have been considering how I would display data from a geo perspective and found your model has helped me to consider exactly how I will build it. Seeing you use the versions so liberally has really made me want to do the same; I've been very frugal with versioning up to now but seeing yours has inspired me to really go for it. I look forward to playing with your model so thanks for the files too.
Kind regards,
Paul Wyatt