Dashboard Potential Enhancements

[WraySmith]

Additional dashboard enhancements that may be included depending on timelines and for discussion with @firasm and @Davddng

• Putting the board game spreadsheet into a database format and calling directly from the database - potentially worth trying to start implementing early on to avoid rewriting code
• Automated updating of the board game database using the scraper available to do so (likely implemented at the very end of the project)
• Doing an unsupervised learning analysis and showing the results in a network graph of similar games (separate tab on the dashboard), potentially showing on a 3D plot using the first 3 principle components or something similar if it works - another at the end item
• Having all or some components of the dashboard AWS based. Initial thoughts include: calculations run using AWS lambda (on demand calcs and would potentially be only applicable to the unsupervised learning analysis), storing the database in dynamoDB or AWS RDS, or implementing full deployment through beanstalk - another in the last week thing going from Heroku to direct AWS hosting

[WraySmith]

I figure the database item is likely the only one that needs to really be considered closer to the beginning but what is done with the database now may impact use/choice of dynamoDB vs RDS if we end up going to full AWS hosting so probably worth thinking about.

[WraySmith]

Discussion points from meeting with @firasm:

• Having the ability to update the dataset on a semi-regular basis should likely be the highest priority enhancement followed by the network analysis (Look into Dash Cytoscape for network graphs #9)
• Likely not a need to put the dataset into a relational database (it's also a pretty flat structure) but could potentially use MongoDB or dynamoDB if looking to do so out of interest
• AWS hosting (Deploying Dash apps on AWS #10) could be looked at after everything is running on Heroku (and probably a better handle at that point of what needs to be done anyways - updating unsupervised analysis using lambda etc)

[WraySmith]

Closing as these have been moved into individual issues #9, #10, #15, #16