georgetown-cset / parat

🦜 PARAT: CSET's Private-sector AI-Related Activity Tracker
https://parat.cset.tech
Other
4 stars 0 forks source link

Group options in patent trends chart's dropdown #387

Closed brianlove closed 1 month ago

brianlove commented 1 month ago

Organize the various options in the dropdown menu controlling the patent trends chart to group related options together and remove other (non-charted) options.

Closes #337, closes #357

github-actions[bot] commented 1 month ago

No need for rebasing :+1: behind_count is 0 ahead_count is 1

github-actions[bot] commented 1 month ago

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines Covered Coverage Threshold Status
594 423 71% 0% 🟢

New Files

No new covered files...

Modified Files

No covered modified files...

updated for commit: 78c516e by action🐍

github-actions[bot] commented 1 month ago

JavaScript Coverage

Summary

Lines Statements Branches Functions
Coverage: 67%
67.45% (398/590) 54.18% (207/382) 68.1% (126/185)
Modified Files • (67%)
File% Stmts% Branch% Funcs% LinesUncovered Line #s
All files67.4554.1868.167.2 
components64.1754.3662.5963.72 
   DetailViewPatents.jsx000015–216
brianlove commented 1 month ago

I don't think the Microsoft in [Robotics, Document Management & Publishing, Military] cases you mentioned are consistent with #357. In that issue, the problem is that there are no underlying data (counts is an empty array), so the x-axis shows the interval [-1, 6] (likely a Plotly default). In these cases, however, the data are present (via the group's counts value in overall_data.json), they are just extremely low (mostly zeros for the y-axis, with one year reaching 1). The x-axis, however, for these cases still have the correct years displayed (2013-2023).

We may want to handle low-data categories like these differently, but I think that's an analytic question to resolve, not a technical issue with the current state of the app.

jmelot commented 1 month ago

Ah ok got it, I misread that issue. I'll ping Zach about what to do about subfields without patent data