Hey <@U0A2E9B5P2N>! I've jot down the feedback on ...

# gsoc-education

[Elizaveta Artser]

Hey @Mark Seif! I've jot down the feedback on the dashboard, it might look a bit scary because of the length but I swear it's not that bad blob sweat smile
I've included some ideas on additional statistics, and some of them would probably require some add-ons in the scrapping pipeline or post-processing, here's a short summary:
• Topics distributions - AI part of the pipeline tagging the records with the associated topics
• Distributions by regions - might be done on post-processing stage as a statical mapping country -> region
• GitHub forks - similar to stars, just as another metric for popularity
• Formal / Informal deduplicated - that's the outcome of the mapping we've discussed previously (mapping GitHub repos to unis / MOOCs)
--- --- --- --- --- --- --- ---

1️⃣ Be careful with types of education (see here for clarity): university courses are part of formal education, MOOCs belong to non-formal education, GitHub repo might be either part of formal edu if accompanying a specific university course, or part of non-formal if standing alone or accompanying specific MOOC

2️⃣ Some terms and metrics need clarification for external users:
• Primary vs Secondary signal
• Formal vs Informal Non-formal learning type
• Kotlin confidence
• Methodology of data scrapping explained before data on search stats
3️⃣ I'd suggest to introduce sections for readability and clarity:
• General
• Formal education (aka universities for now)
• MOOCs
• GitHub
• Search statistics
General:
• What we already have:
◦ Total records, Unique universities, Unique countries, GitHub repos, MOOCs - in numbers
◦ By source (University courses / MOOCs / GitHub repos)
◦ Formal / Informal
◦ Primary / Secondary signal
• What would be nice to add:
◦ Formal / Informal deduplicated (map GitHub repos to university courses / MOOCs and take mapped repos out of consideration)
◦ Topics distribution
Formal education:
• What we already have:
◦ Map of universities teaching Kotlin by country
◦ Countries with most universities teaching Kotlin (top-15)
• What would be nice to add:
◦ Universities with most courses teaching Kotlin (top-15)
◦ Universities teaching Kotlin, distribution by region
◦ Courses teaching Kotlin, distribution by region
◦ Courses teaching Kotlin / Other resources and mentions on Kotlin
◦ Topics distribution in formal education, overall
◦ Topics distribution by region
MOOCs:
• What we already have:
◦ Distribution by platform
• What would be nice to add:
◦ Topics distribution
◦ List of all courses
GitHub:
• What we already have:
◦ Repository type (course / workshop / tutorial / book companion / library or app / personal learning)
◦ GitHub stars distribution
◦ Top providers by number of repos (top-15)
• What would be nice to add:
◦ Association with other learning formats (associated with known university course / associated with known MOOC / stand-alone or not mapped)
◦ GitHub forks distribution
◦ Top repos by stars (top-15)
◦ Top repos by forks (top-15)
Search statistics:
• Already good enough, I'd only suggest to use smaller buckets and center bars on horizontal ticks in the bar chart for Kotlin-confidence distribution (now it reads confusing)
• Also, as said above, it would be nice to briefly describe the methodology of data scrapping and make sure the terms are consistent across methodology and charts
4️⃣ Some UI issues noticed:
• Filters are confusing - if activating the filter includes this category in the search and on the dashboard open all records are presented, then all corresponding filters should be already pre-activated (this might be also causing inconsistencies in filtering)
• Primary-only filter sounds obsolete, given the Primary and Secondary as separate filters
• Sliders seem not the best choice for lower and upper bound of Kotlin confidence, I'd suggest using two number input fields instead for preciseness
• Not sure whether source filter is feasible, it's only meaningful in the general section, wdyt?
• GitHub stars filter, given that we already have GitHub stars distribution chart, seems unnecessary, wdyt?
• Filtering from the dashboard itself is only functional for countries, is this correct? If so, I'd state this functionality somewhere in the description
• Not all names are displayed neatly - some are already shortened with ... but also the name's beginning is not displayed fully (see a screenshot)
• The data table can be scrolled horizontally - but the heading row is not scrolled with the data
• If that's possible, I'd suggest to include the date-time of either export of the dataset from the website or of the last update of the dataset
--- --- --- --- --- --- --- ---

Hope these comments will be helpful 🤞 Don't hesitate to ask any questions or discuss any of the suggestions!
And just to make it clear, I still believe the dashboard is amazing and you did a great job so far! 💪

@Ksenia Shneyveys please also comment if you have any suggestions!