I'll take more detailed look at this tomorrow.

Noting for myself that the forecast page currently does not have the new clades:

This is likely because the sequences counts do not reach the clade_min_seq config

We can force include the new clades with something like (untested):

diff --git a/config/config.yaml b/config/config.yaml
index f0a2307..787ce8e 100644
--- a/config/config.yaml
+++ b/config/config.yaml
@@ -22,6 +22,7 @@ prepare_data:
         prune_seq_days: 12
         clade_min_seq: 2000
         clade_min_seq_days: 150
+        force_include_clades: "24H=24H 24I=24I"
     pango_lineages:
       global:
         included_days: 150

I think we should lower thresholds in general. There are many interesting countries with relevant sequencing activity that are excluded. I did a test run with lower thresholds across the board and it looks good. Will share.

Sequencing activity keeps dropping, so without dropping thresholds the forecasts become increasingly irrelevant.

Same for pango lineages. LP.8 has around 140 sequences now. It's growing fast, it's been noteworthy for multiple weeks but will show only in a few weeks with current thresholds demanding 350 sequences.

I can make a proposal tomorrow.